Multiplex CRISPR-Cas9 knockout of EIL3, EIL4, and EIN2L advances soybean flowering time and pod set

Abstract

Background

Ethylene inhibitor treatment of soybean promotes flower bud differentiation and early flowering, suggested that there is a close relationship between ethylene signaling and soybean growth and development. The short-lived ETHYLENE INSENSITIVE2 (EIN2) and ETHYLENE INSENSITIVE3 (EIN3) proteins play central roles in plant development. The objective of this study was carried out gene editing of EIL family members in soybeans and to examine the effects on soybean yield and other markers of growth.

Methods and results

By editing key-node genes in the ethylene signaling pathway using a multi-sgRNA-in-one strategy, we obtained a series of gene edited lines with variable edit combinations among 15 target genes. EIL3, EIL4, and EIN2L were editable genes favored by the T0 soybean lines. Pot experiments also show that the early flowering stage R1 of the EIL3, EIL4, and EIN2L triple mutant was 7.05 d earlier than that of the wild-type control. The yield of the triple mutant was also increased, being 1.65-fold higher than that of the control. Comparative RNA-seq revealed that sucrose synthase, AUX28, MADS3, type-III polyketide synthase A/B, ABC transporter G family member 26, tetraketide alpha-pyrone reductase, and fatty acyl-CoA reductase 2 may be involved in regulating early flowering and high-yield phenotypes in triple mutant soybean plants.

Conclusion

Our results provide a scientific basis for genetic modification to promote the development of earlier-flowering and higher-yielding soybean cultivars.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-023-04543-x.

Background

Ethylene plays an important regulatory role in plant growth and development, particularly in response to biotic and abiotic stress. A linear model of ethylene signal transduction has been established by screening the ethylene "triple response" mutants [1]. In this model, CTR1 is a negative regulator of the ethylene signal transduction pathway and phosphorylates EIN2 after binding with the receptor. The phosphorylation-regulated proteolytic processing of EIN2 triggers its trafficking from the endoplasmic reticulum (ER) to the nucleus. Ethylene triggers dephosphorylation at several sites and proteolytic cleavage at one of these sites, resulting in nuclear translocation of a carboxyl-terminal EIN2 fragment (EIN2-C'). The EIN2-C' peptide combines with EIN3 to jointly regulate the downstream target genes [2], which further activates downstream ethylene-related transcription factors, such as ERF1, and amplifies the ethylene signal [3]. In the ethylene signal transduction pathway, the EIN2-C' peptide combines with transcription factors ETHYLENE INSENSTIVE 3 (EIN3) and ETHYLENE INSENSITIVE 3-LIKE 1 (EIL1) to positively regulate the ethylene signaling pathway [2].

EIN2 and EIN3/EIL have been identified in many plants. For instance, EIN2 and EIN3/EIL affect seedling growth and responses to salt and other stressors in Arabidopsis [4–7]. In Petunia, reduced PhEIN2 expression delays flower senescence and fruit ripening, and reduces adventitious root and seedling root-hair formation [8]. In rice, homologous genes of Arabidopsis ethylene signaling components have been identified, including ETR, RTE1-like, EIN2-like, and EIN3-like genes, which control grain size and weight, root growth, salt tolerance, and wound signaling [9–12]. In soybean, comparative transcriptome analysis has shown that exogenous ethephon treatment might activate EIN3 and ERF1/2 in the ethylene signal transduction pathway, amplify ethylene signaling, inhibit soybean flowering, and promote flowering reversal. Ethylene inhibitor treatment of soybean inhibits the transcriptional activity of ERF1/2, promotes flower bud differentiation, and leads to early flowering [13–15]. Although genome-wide analyses of the EIN2 and EIN3/-EIL gene family has been carried out for several plant species, no comprehensive analysis of this family has been reported for soybean. Since the soybean is a globally important agricultural crop, the EIL gene family should be thoroughly investigated in its biology.

To characterize the function of EIN2 and EIN3/EIL gene family members in plant development, the mutation phenotype and overlapping, opposite or different functions of EIN2 and EIN3/EIL gene family members should be analyzed. The ethylene-response phenotype in Arabidopsis ein3 mutant was lost due to the reduced expression of ERF1 and chitinase genes but the mutant phenotype could be rescued by the overexpression of EIL1 or EIL2, which indicates that EILs play an important regulatory role in soybean ethylene signaling pathway [16, 17]. LeEILs members are functionally redundant and played vital role in regulating ethylene reaction during the whole life cycle of tomato. However, genetic mutant screens of EIL family members in soybean are limited due to its paleopolypoid genome. Another major obstacle in the gene editing application of CRISPR-Cas9 for soybean is its low transformation efficiency. The key components in the ethylene pathway contain several gene family members, and, therefore, knocking out subjectively and randomly only a single gene would lead to weaker phenotypic differences. Fortunately, multiplex mutations can now be generated by either cloning multiple sgRNAs or targeting multiple homologous regions by a single sgRNA in the CRISPR-Cas9 system in both dicot and monocot plants [18].

In our study, one pYLCRISPR-Cas9 plasmid with 15 sgRNA cassettes for editing 15 EIN2 and EIN3/EIL genes was constructed for soybeans guided by natural variation. The screening of mutants followed by cultivation experiments provided useful information for the selection of EIN2 and EIN3/EIL genes that are agronomically important.

Results

Phylogenetic analysis of EIN2 and EIL

A total of 260 EIN2s and 1425 EILs were found in Ensembl Plants (http://plants.ensembl.org/) respectively. There were three EIN2s and eight EILs in genome of Arabidopsis thaliana. We identified 12 EIL and four EIN2 genes in the soybean W82 genome. Compared to Arabidopsis, soybean has more EIN2s and EILs members, which is consistent with the more complex paleopolypoid genome of soybean. To explore the phylogenetic relationships of the EIN2 and EIL families, a neighbor-joining (NJ) tree was constructed by aligning the sequences with 89 other sequences from A. thaliana, Cicer arietinum, Glycine soja, Lotus japonicus, Medicago truncatula, Oryza sativa, Phaseolus vulgaris, Prunus persica, Salix purpurea, and Vitis vinifera (Fig. 1A and B). The results of systematic evolution analysis showed that the 21 EIN2s and 84 EILs from different plants were divided into two and four evolutionary branches, respectively (Fig. 1A and B). Four EIN2 genes belonged to two evolutionary branches, indicating that they may come from the two ancestral genes (Fig. 1A). All EILs were divided into four evolutionary branches, each with five, one, two, and four members, indicating that these 12 EIL genes may come from four ancestral genes (Fig. 1B). The results also showed that most of 12 EIL and four EIN2 gene members from G. max had a relatively close genetic relationship with G. soja, L. japonicus, and P. vulgaris, while had a relatively distant genetic relationship with A. thaliana and O. sativa.

Fig. 1.

Phylogenetic tree of EIN2 and EILs proteins in plants. A Phylogenetic tree of EIN2s in soybean cultivar Williams 82. B Phylogenetic tree of EILs in soybean cultivar Williams 82. Red font indicates 12 EILs and 4 EIN2 proteins in soybean cultivar Williams 82

CRISPR-Cas9 system for targeted mutagenesis of EIN2 and EIL genes in soybean

The CRISPR-Cas9 vector system for multiplex targeting of genomic sites was used to clone the target sequences of 15 genes into one vector and carry out genetic transformation. The whole genomic DNA of transformed plants was analyzed by PCR and sequencing. Combining with PAT/Bar test strip detection, 83 T-DNA positive plants were obtained in the T0 generation, and further sequencing of the gene targets in the transformed plants showed that 13 genes targets were successfully edited in the 74 T0 generation lines. All the 13 target genes were successfully knocked out in one plant, which caused arrested growth and development (Fig. 2A and B). Among the 83 glufosinate-resistant T0 plants, two, three, and four SNP mutations had the highest frequencies in 18, 20, and 21 plants, accounting for 71.08% of the total glufosinate-resistant T0 plants (Fig. 2B). Among all mutated plants, the number of plants without indel mutations was the highest, accounting for 39.76% of the total; among plants with indel mutations, the number of plants with two or three indel mutations was the highest, accounting for 54.22% of the total (Fig. 2C). Sequencing the 13 target genes in 74 positive T0 plants confirmed simultaneous editing of EIL3 (Glyma.15G031800), EIL4 (Glyma.18G018400), and EIN2L (Glyma.10G058300) in 23 lines. This suggests that these three genes are preferred editable genes in soybean, and their simultaneous editing may be conducive to cell reproduction.

Fig. 2.

Comparative assays of single-nucleotide polymorphism/insertion or deletion (SNP/InDel) in the T0 soybean using the CRISPR/Cas9 gene editing technology. A Single-nucleotide polymorphisms and haplotype analysis in EILs and EIN2 genes in different T0 lines. B Number of characterized SNP mutant in T0 gene editing events. C Number of the characterized indel mutant in T0 gene editing events

EIL3, EIL4 and EIN2L triple mutants

In EIN2L, EIL3, and EIL4 triple mutant lines (abbreviated as experiment number Z4) of T2 generation, we identified Cas9-free plants carrying targeted mutations in which the T-DNA was lost while the EIN2L, EIL3, and EIL4 genes were simultaneously edited. Sequencing of the EIL3, EIL4, and EIN2L T3-mutant plants revealed a 10-bp deletion at T1 + sgRNA-targeted EIL3, a 4-bp deletion at T2 + sgRNA-targeted EIL4, and a 5-bp deletion at T3 + sgRNA-targeted EIN2L (Fig. 3A and B), resulting in frameshift mutations in the EIL3, EIL4, and EIN2L proteins (Fig. 3C). The editing sites of EIL3 and EIL4 were at the 5' end of the gene, indicating that most of the amino acids in their proteins were mutated from the N' end. The editing sites of EIN2L were biased towards the 3' end, indicating that the EIN2-C' peptide of its protein was altered.

Fig. 3.

Soybean triple mutation of EIL3, EIL4 and EIN2L generated by the CRISPR/Cas9 system in the T3 generation. A Sequences of WT and mutant plants at target sites EIL3-sg-T1, EIL4-sg-T2, and EIN2L-sg-T3. Dashes indicate deleted nucleotides. Proto-spacer adjacent motif (PAM) has been highlighted with red letters. B The sequence peaks of the wild-type (WT) and mutants at target sites EIL3-sg-T1, EIL4-sg-T2, and EIN2L-sg-T3. Red arrows indicate mutation locations. C Amino acid sequences of the triple mutation of EIL3, EIL4 and EIN2L and WT

Phenotypic changes in triple mutant plants

Compared to that of the wild-type control, the early flowering stage R1 of Z4 was 7 d earlier (Figs. 4A and 5A), the early podding stage R3 of the triple mutant was 6 d earlier (Figs. 4B and 5B), the full pod stage R4 of the triple mutant was 8 d earlier (Fig. 5C), and the maturation stage R8 of the triple mutant was 7 d earlier (Figs. 4C and 5D). At the maturity stage R8, all leaves were removed and the differences in pod-setting characteristics between the two treatments were compared. Compared to the Z4 plants, fewer pods formed on the top branches of the control plants (Fig. 4D). Furthermore, the average number of pods per plant was 133.65 for the Z4 plants compared to 81.18 for the wild-type controls (Figs. 4E and 5E). At harvest, an average of 293.29 beans were harvested per Z4 plant compared to 164.69 for each control plant, and the number of seeds per plant was also 1.78-fold higher in the Z4 plants (Fig. 5F). The average 100-grain weight of the Z4 plants was 17.89 g compared to 19.28 g for the wild type (Fig. 5G), and the yield of Z4 plant was 52.67 g compared to 31.85 g for the wild-type plant (Figs. 4F and 5H). Overall, the EIL3, EIL4, and EIN2L triple mutant had a significantly accelerated growth rate, improved pod-setting at the top of the plants, and significantly increased yield by 65% compared with the wild-type control plants.

Fig. 4.

Phenotypes of EIL3, EIL4 and EIN2L triple mutant Z4 and WT. A Triple mutation leads to early flowering; B Triple mutation leads to early podding; C Triple mutation leads to premature pods; D Triple mutation plants and control at harvest; E Pods of triple mutation plants and control; F Seeds of triple mutation plants and control. WT, wild type; Z4, triple mutation plant. Z4-1, -2 and -3 are three replicate plants from the Z4 triple mutant line

Fig. 5.

Growth period and pod setting habit changes of triple mutant. A Flower stage changes; B Pod stage changes; C Full pod stage changes; D Mature stage; E Changes of pods per plant; F Changes of seeds number per plant; G changes of 100 grain weight; H changes of yield per plant. DAE, days after emergence. CK, WT of soybean W82; Z4, triple mutation plant. All physiological indexes were measured for at least 10 individual plants. Means were compared using LSD (least significant difference) t-test at the 5% level of significance. ** represents significant difference at P ≤ 0.05 as indicated by LSD test

Identification of differentially expressed transcripts

After sequencing data quality control, six digital gene expression (DGE) libraries were generated composed of 169,592,355 clean reads covering a length of 50,498,587,958 clean bases (Table S1). Each clean read was 298-bp long. The obtained clean reads were mapped to the W82 genome, and the mapping ratio ranged from 95.63% to 96.16% (Table S2). All raw transcriptome data were deposited in the sequence read archive (accession no. PRJNA774476). The mapped reads were assembled and aligned to the reference genome to obtain annotations. Differentially expressed genes (DEGs) were identified using DESeq2 according to a fold change (FC) ≥ 2 and false discovery rate (FDR) < 0.01. Compared to the WT, 412 upregulated and 461 downregulated unigenes were identified in the triple mutant (Fig. 6; Table S3 and S4).

Fig. 6.

Volcano map of differentially expressed genes between control and triple mutation plants. Note: Each point in the volcano map represents a gene, and the abscissa represents the logarithm of expression fold change of a gene in the two samples. The ordinate represents the negative logarithm of the statistical significance of the change in gene expression. The greater the absolute value of abscissa, the greater the difference of expression multiple between the two samples. The larger the ordinate value, the more significant the differential expression, and the more reliable the differentially expressed genes screened. A fold change (FC) ≥ 2 and false discovery rate (FDR) < 0.01 was used to identify DEGs. Up or down-regulated genes refer to those in the Z4 triple mutant relative to WT control plants. Green dots represent down regulated differentially expressed genes, red dots represent up regulated differentially expressed genes, and black dots represent non differentially expressed gene (non-DEG)

KEGG enrichment analysis of differentially expressed genes

A total of nine KEGG pathways were significantly enriched, with Q-values smaller than 0.05 (Fig. 7; Table S5). These DEGs were associated with linoleic acid metabolism (ko00591); starch and sucrose metabolism (ko00500); cutin, suberine, and wax biosynthesis (ko00073); phenylpropanoid biosynthesis (ko00940); diterpenoid biosynthesis (ko00904); ABC transporters (ko02010); alpha-linolenic acid metabolism (ko00592); flavonoid biosynthesis (ko00941); and sesquiterpenoid and triterpenoid biosynthesis (ko00909). Most of the significantly enriched KEGG pathways were related to metabolism and biosynthesis, suggesting that the expression of these genes in the Z4 flower organ may be responsible for its earlier flowering and high-yielding phenotype.

Fig. 7.

Significantly enriched pathways in control and triple mutation plants paired comparisons. Note: P-value ≤ 0.05 is set as the significant threshold for significantly enriched GO terms

DEGs in plant hormone signal transduction pathways

EIL3, EIL4, and EIN2L are important gene members in ethylene signal transduction and their knockout may have important impacts on plant hormone signaling due to crosstalk effects. The editing of these three genes had no significant effect on ethylene signaling pathway (Figure S1; Table 1). In the triple mutant, nine genes involved in auxin signaling were up-regulated 1 to 1.8-fold (Figure S1; Table 1). This included two AUX1 genes (Glyma.03G063900 and Glyma.13G319700), five Aux/IAA genes (Glyma.02G142400, Glyma.07G018100, Glyma.07G034200, Glyma.19G161100, and Glyma.20G210400) and two SAUR genes (Glyma.08G155700 and Glyma.12G034100). In the cytokinin signal transduction pathway, a DEG (Glyma.03G241700) encoding AHP was upregulated by 1.03-fold, and two DEGs (Glyma.05G158900 and Glyma.08G116700) encoding B-ARR were downregulated by 1.12-fold and 1.17-fold, respectively (Figure S1; Table 1). In the jasmonic acid signal transduction pathway, four JAZ genes (Glyma.01G204400, Glyma.11G038600, Glyma.13G112000, and Glyma.13G116100) and MYC2 (Glyma.15G166800) was also downregulated by 1.14 to 2.50-fold (Figure S1; Table 1). In the salicylic acid signal transduction pathway, a TGA6 (Glyma.09G184300) was downregulated by 4.93-fold (Figure S1; Table 1). In the brassinosteroid signal transduction pathway, five BRI1 genes (Glyma.06G184400, Glyma.18G254000, Glyma.18G254500, Glyma.18G254600, and Glyma.19G081200) were upregulated by 1.22 to 2.08-fold, while SBP13 and SBP1 (Glyma.11G083900 and Glyma.13G177400) were downregulated by 1.25 and 1.24-fold (Figure S1; Table 1). Based on these results, gene editing of EIL3, EIL4, and EIN2L may influence the signal transduction pathways of auxin, cytokinin, jasmonic acid, salicylic acid, and brassinosteroids.

Table 1.

DEGs in plant hormone signal transduction pathways

Gene ID	FDR	Log2FC	Annotation	Signaling pathway
Glyma.03G063900	2.14E-13	1.04	AUX1	auxin
Glyma.13G319700	2.28E-11	1.07	AUX1	auxin
Glyma.02G142400	3.69E-07	1.07	Aux/IAA	auxin
Glyma.07G018100	1.65E-05	1.01	Aux/IAA	auxin
Glyma.07G034200	1.99E-13	1.00	Aux/IAA	auxin
Glyma.19G161100	6.50E-120	1.80	Aux/IAA	auxin
Glyma.20G210400	3.57E-45	1.33	Aux/IAA	auxin
Glyma.08G155700	1.07E-05	1.03	SAUR	auxin
Glyma.12G034100	3.66E-05	1.03	SAUR	auxin
Glyma.03G241700	1.28E-05	1.03	AHP	cytokinin
Glyma.05G158900	1.94E-17	-1.17	B-ARR	cytokinin
Glyma.08G116700	4.48E-17	-1.12	B-ARR	cytokinin
Glyma.01G204400	3.29E-51	-2.50	JAZ	jasmonic acid
Glyma.11G038600	9.66E-48	-1.97	JAZ	jasmonic acid
Glyma.13G112000	4.38E-24	-1.14	JAZ	jasmonic acid
Glyma.13G116100	4.23E-104	-1.42	JAZ	jasmonic acid
Glyma.15G166800	3.06E-08	-1.20	MYC2	jasmonic acid
Glyma.09G184300	5.03E-137	-4.93	TGA6	salicylic acid
Glyma.06G184400	3.94E-19	2.08	BRI1	brassinosteroid
Glyma.18G254000	3.51E-09	1.42	BRI1	brassinosteroid
Glyma.18G254500	1.28E-10	1.53	BRI1	brassinosteroid
Glyma.18G254600	5.28E-08	1.27	BRI1	brassinosteroid
Glyma.19G081200	2.54E-21	1.22	BRI1	brassinosteroid
Glyma.11G083900	1.14E-08	-1.25	SBP13	brassinosteroid
Glyma.13G177400	5.57E-20	-1.24	SBP1	brassinosteroid

The character "glyma.Wm82.gnm2." are omitted before gene IDs. FDR, false discovery rate; FC, fold change: FPKM of triple mutants divided by FPKM of control; AHP, histidine-containing phosphotransfer protein; Aux/IAA, auxin-responsive protein IAA; AUX1, auxin influx carrier; B-ARR, two-component response regulator ARR-B family; BRI1, protein brassinosteroid insensitive 1; JAZ, jasmonate ZIM domain-containing protein; MYC2, transcription factor MYC2; SAUR, SAUR family protein; SBP, squamosa promoter binding like protein; TGA, transcription factor TGA. The complete annotations for each DEG are listed in Table S4

Validation of DEGs and non-DEGs by qRT-PCR

Six DEGs and six negative control genes were selected for the qRT-PCR analysis. Among all the non-DEGs, five were expressed in the terminal bud at the flowering stage (EIL3, EIN2L, XTH23, LNK3, and IMP4), and both the qRT-PCR and RNA-seq data suggested that the RNA-seq and qRT-PCR data were well correlated (Fig. 8). Both EIN3 and EIN2L were expressed in the terminal bud, but their expression did not change significantly. In contrast, EIL4 was not expressed in the terminal bud (Table S3 and Table S6) and was considered a non-constitutively expressed gene. Among the six DEGs (SS, ABCG26, PKSA, AGL42, FAR2, and AGAMOUS), RNA-Seq and qRT-PCR confirmed significant expression differences between wild-type and Z4 plants. Taken together, the expression patterns from the qRT-PCR analysis were consistent with those obtained from the RNA sequencing.

Fig. 8.

Validation of non-DEGs and DEGs by qRT-PCR analysis. Bars represent mean ± standard deviation (n = 3). The significance level was set as P < 0.05. Note: EIL3, Ethylene insensitive 3; EIN2L, Ethylene insensitive 2 like; XTH23, xyloglucan endotransglucosylase/hydrolase protein 23; LNK3, Protein LNK3; IMP4, U3 small nucleolar ribonucleoprotein protein IMP4; SS, Sucrose synthase; ABCG26, ABC transporter G family member 26; PKSA, Type III polyketide synthase A; AGL42, MADS-box protein AGL42; FAR2, Fatty acyl-CoA reductase 2; AGAMOUS, Floral homeotic protein AGAMOUS

Discussion

Cell signaling is a complex process involving many gene networks that is characterized by the non-discrete and non-linear flow of information. Many genes belong to larger gene families and, like soybeans, crops displaying paleopolyploidy have several copies of each gene. Genome complexity leads to high gene redundancy in soybean, which limits the output of mutant screening by forward genetic approaches [19]. A large-scale CRISPR-Cas9 mutagenesis population would be of great value for crop research, and recent studies have generated mutant populations in soybean via pooled transformation of CRISPR-Cas9 libraries [20–22]. To the best of our knowledge, we are the first to edit key-node genes in the ethylene signaling pathway using a multi-sgRNA-in-one strategy. We successfully generated a series of gene edited lines with altered combinations of 15 edited gene targets. We found that two sgRNAs failed to induce mutations in all the T0 plants, six sgRNAs induced low-frequency mutation (< 10%), and one sgRNA induced mutation with a frequency of 67.5%. Although computational models based on DNA sequence have been used to predict the efficiency of sgRNAs, the actual performance of sgRNAs within cells varies greatly [23, 24]. The specific gene editing mutation are often greatly influenced by the efficiency of Cas9, the status of host DNA repair machinery, the DNA damage type, and the structure of DNA repair donor molecules [25]. In the ethylene-insensitive mutant lines of T2 generation, we identified Cas9-free plants carrying targeted mutations, in which the T-DNA was lost while the EIN2L, EIL3, and EIL4 genes were all edited. This suggests that the sgRNAs with high frequencies were favored in the transformed plants, and their simultaneous editing may be conducive to the growth and development of soybean.

Ethylene is a stress hormone that temporally accelerates organ abscission in many plant species [26, 27]. Several key components of ethylene signaling, such as ETR1 and EIN2, have been found to induce floral organ shedding in Arabidopsis, but the ethylene-insensitive (ETR1-1, EIN2-1) mutant lines significantly delayed shedding in Arabidopsis [26–28]. In soybean, treatment with the ethylene inhibitor aminoethoxyvinylglycine increases soybean yield by reducing flower pod shedding [15]. These results suggest that modulating ethylene biosynthesis and signal transduction pathways may be an effective method for creating high-yield germplasm. Following gene editing, we found that EIL3, EIL4, and EIN2L triple mutants expressed earlier flowering (by approximately 7 d) and high-yield (a 65% increase) phenotypes when compared to the wild type plants. Unexpectedly, EIL4 was not expressed in the terminal bud. In the soybean transcriptome database (https://ngdc.cncb.ac.cn/soyomics/transcriptome/) [29], it was found that EIL4 (Glyma.18G018400) was mainly expressed in seeds during the mature stage R8, suggesting its potential roles in regulation of seed maturation, which was consistent with our result that EIL4 was not expressed in the terminal bud. Similarly, we also found spatiotemporal expression features of EIL3 (Glyma.15G031800) and EIN2L (Glyma.10G058300) in the aforementioned database. EIL3 is specifically highly expressed in floral organs during flowering stage, and this gene is also expressed in pods. EIN2L is mainly expressed in floral organs before flowering, as well as in pods and young seeds. Based on this, it is inferred that EIL3 and EIN2L are involved in the regulation of floral organ, pod, and seed development in soybean. Therefore, the simultaneous editing of these three genes may achieve high yield through regulation of floral organ development, pod development, seed development, and seed maturation. EIL3 and EIN2L are expressed in floral organs at different stages, indicating that the editing of them genes may be related to the early flowering of mutant plants.

Comparative transcriptome analysis of the triple mutant and wild type plants revealed 873 DEGs. The number of DEGs was highest in the starch and sucrose metabolism pathway, which contained 33 DEGs. Sucrose synthase (SS), which produces sucrose and uridine diphosphate (UDP), is closely related to flowering and flower development; photosynthetic products mainly transported in the form of sucrose. In pepper (Capsicum spp.), sucrose transport into the flowers enlarges the flower sink by suppressing abscission and activating metabolism, resulting in a higher rate of flower set [30]. Furthermore, in cotton (Gossypium spp.), GhSusA1 suppression in transgenic plants decreases boll size and seed weight [31]. Based on our results, the expression of 11 DEGs encoding SS in soybean was upregulated by 1.30–4.52 fold compared to nine genes encoding SS that were downregulated by 1.83–3.38-fold (Table 2; Table S4). Overall, compared with the downregulated genes, the amount, expression abundance, and FC of the upregulated genes were dominant. Thus, we speculate that sucrose synthesis and the supply of photoassimilates to axillary buds are enhanced in the triple-mutant plants, which is conducive to flower organ development and increased yield.

Table 2.

Genes of interest that may be involved in the regulation of flower / seed development

Gene ID	FDR	Log2FC	Annotation
Glyma.04G180300	2.92E-258	6.12	At2g26730
Glycine_max_newGene_3783	3.06E-227	5.89	L9
Glyma.03G054100	2.94E-113	4.52	SS
Glyma.03G053500	1.41E-80	3.99	SS
Glyma.02G302200	4.52E-48	2.73	ABCG26
Glyma.11G097900	5.62E-40	2.63	PKSB
Glyma.07G043300	1.89E-31	2.43	CYP704B1
Glyma.01G073600	1.12E-74	2.41	PKSA
Glyma.14G012000	1.56E-26	2.16	ABCG26
Glyma.08G332100	1.48E-20	2.16	ABCG26
Glyma.15G018500	9.43E-37	2.09	TKPR
Glyma.09G104800	3.06E-19	2.08	FAR2
Glyma.06G184400	3.94E-19	2.08	BRI
Glyma.19G161100	6.50E-120	1.80	AUX28
Glyma.20G210400	3.57E-45	1.33	IAA14
Glyma.10G240500	2.29E-07	1.24	AGL42
Glyma.11G252300	1.59E-06	1.14	MTF1
Glyma.15G058000	4.71E-15	1.06	JOINTLESS
Glyma.13G223300	5.56E-09	1.04	AGAMOUS
Glyma.07G081300	1.71E-06	1.02	MADS3

Except for gene Glycine_max_newGene_3783, the character "glyma.Wm82.gnm2." are omitted before other gene IDs. FDR, false discovery rate; FC, fold change: FPKM of triple mutants divided by FPKM of control; At2g26730, inactive receptor kinase At2g26730; L9, 50S ribosomal protein L9; SS, Sucrose synthase; ABCG26, ABC transporter G family member 26; PKSB, Type III polyketide synthase B; CYP704B1, Cytochrome P450 704B1; PKSA, Type III polyketide synthase A; TKPR, Tetraketide alpha-pyrone reductase; FAR2, Fatty acyl-CoA reductase 2; BRI, Brassinosteroid Insensitive 1; AUX28, Auxin-induced protein AUX28; IAA14, Auxin-responsive protein IAA14; AGL42, MADS-box protein AGL42; MTF1, MADS-box transcription factor 1; JOINTLESS, MADS-box protein JOINTLESS; AGAMOUS, Floral homeotic protein AGAMOUS; MADS3, Agamous-like MADS-box protein MADS3. The complete annotations for each DEG are listed in Table S4

The MADS-box genes encode important transcription factors important for flower and fruit development in plants. Significantly, four soybean MADS-box genes were > onefold up-regulated in the triple mutant and their orthologs in other plants have critical roles in flowering and fruit development. Glyma.10G240500 encodes AGL42 (Table 2; Table S4) and inactivation of its Arabidopsis ortholog (AGL42) produces a non-flowering phenotype [32]. Glyma.15G058000 encodes JOINTLESS (Table 2; Table S4) and inactivation of its Solanum spp. ortholog prevents abscission zone formation [33]. Glyma.13G223300 encodes AGAMOUS (Table 2; Table S4) and inactivation of its Arabidopsis ortholog (AGAMOUS) decreases the number of mature ovules and increases primordial outgrowth [34]. Glyma.07G081300 encodes MADS3 (Table 2; Table S4) and its grapevine (Vitis vinifera L.) ortholog is specifically expressed in both male and female flowers and in developing seeds [35]. This suggests that MADS-box genes may affect the yield and flowering period in the triple mutants by regulating flower development.

In maize (Zea mays), kernel 44 (DEK44) encodes a putative 50S ribosomal protein L9; DEK44 mutants produce small kernels with embryo-lethal phenotypes [36]. We found that a DEG (Glycine max newGene_3783) encoding 50S ribosomal protein L9 was highly upregulated by 5.89-fold in the triple-mutant soybean plants (Table 2; Table S4). This suggests that 50S ribosomal protein L9 may affect the yield in the triple mutants by regulating abscission zones and kernels.

Comparative proteomic analysis of NJCMS2A (soybean cytoplasmic nuclear male sterile line) and NJCMS2B (corresponding maintainer line) at the two-cell pollen stage showed that four proteins were missing in the anthers of NJCMS2A that were present in NJCMS2B, including MADS-box proteins and auxin-inducible protein AUX28. Therefore, AUX28 may play an important role in the regulation of soybean fertility [37]. In cucumber (Cucumis spp.), the auxin-responsive gene IAA14 is involved in sugar-induced parthenocarpic fruit formation [38]. In our study, five DEGs encoding auxin-responsive IAA were found, all of which were upregulated. Among these, Glyma.19G161100 encodes auxin-induced AUX28, with the highest expression abundance, and was upregulated by 1.80-fold. IAA14 (Glyma.20G210400) was also upregulated by 1.33-fold (Table 2; Table S4). Thus, we speculate that the upregulated expression of AUX28 and IAA14 enhances soybean fertility.

Flowering is a major event in plant development and is regulated by a series of complex signal transduction pathways. Studies have shown that the Brassinosteroid Insensitive 1 (BRI) mutant bri1-5 displays delayed flowering [39]. BRI is dependent on BR signaling and can promote flowering by inhibiting the expression of FLOWERING LOCUS C (FLC) [40]. We identified a total of seven DEGs encoding BRI in soybean, of which five (Glyma.06G184400, Glyma.18G254000, Glyma.18G254500, Glyma.18G254600, and Glyma.19G081200) were upregulated by 1.22–2.08 fold, and two (Glyma.11G083900 and Glyma.13G177400) were downregulated by 1.24–1.25 fold compared to wild-type plants; we found no DEGs encoding FLC (Table 2; Table S4). In Arabidopsis, plants with the RLK6 (inactive receptor kinase At2g26730) overexpression mutation bloom earlier than their wild type counterparts [41]. In our study, two DEGs (glyma.04g180300 and glyma.06g184400) encoding At2g26730 orthologs were upregulated by 6.12-fold and 2.08-fold, respectively (Table 2; Table S4). Thus, we speculate that the upregulated expression of BRI and At2g26730 in the triple-mutant soybean activated the BR signal transduction pathway, resulting in earlier flowering when compared to wild type plants.

The exine is the outer wall of pollen and acts as a protective barrier for the male gametophyte. It also mediates pollen-stigma adhesion. In Arabidopsis, POLYKETIDE SYNTHASE A (PKSA) and POLYKETIDE SYNTHASE B (PKSB) are specifically and transiently expressed in the tapetal cells of anthers during microspore development. Mutants compromised in PKS gene expression display pollen exine layer defects, and a double PKSA PKSB male mutant was found to be completely sterile, with no apparent exine [42]. In rapeseed (Brassica napus), BnPKSA, BnPKSB, and BnACOS5 can form a heterodimer multienzyme complex in tapetal cells that is important for male reproductive development [43]. In Arabidopsis, cytochrome P450 (CYP704B1) [44], ABCG26 [45], TETRAKETIDE alpha-PYRONE REDUCTASE (TKPR) [45], FAR2 (FATTY ACID REDUCTASE 2) [46], ABCG26 [47, 48] are essential for exine development, and the mutations in these genes cause several phenotypes spanning loss of or an aberrant the exine [44, 46] and/or severely reduced fertility [46–48]. In our studies, PKSA, PKSB, CYP704B1, three ABCG26, one TKPR (Glyma.15G018500), and FAR2 genes of soybean were over two-fold upregulated in the triple mutant relative to wild-type plants; in addition, a second soybean TKPR gene (Glyma.07G157200) was upregulated 1.4-fold (Table 2). These data strongly suggest that these soybean genes may help to promote biosynthesis of the pollen exine layer and overall fertility.

Conclusions

For ethylene signaling, the short-lived EIN2 and EIN3 proteins play central roles. To investigate whether differential responses are caused by differential expression of members of the EIN2 and EIN3 family transcription factors, all the family members of soybean were identified. By editing key-node genes in the ethylene signaling pathway using CRISPR/Cas9, we obtained a series of gene edited lines with variable combinations of 15 edited target genes. EIL3, EIL4, and EIN2L mutations were prevalent in the T0 gene edited soybean lines. Here, we characterized a T3 soybean line harboring mutations in EIL3, EIL4, and EIN2L (a triple mutant). The T3 generation of the triple mutants had an early flowering time and high-yield phenotype. The yield of the triple mutants was 1.65-fold higher than that of the wild-type controls. We also compared the transcriptomes of the triple-mutant and wild-type terminal buds and identified 873 DEGs. Based on a functional analysis of the main DEGs reported in the literature, a set of DEGs likely involved in the regulation of the early flowering and high-yield phenotype of the triple-mutant line was identified, including SS, AUX28, IAA14, GIBBERELLIN 20 OXIDASE, BRI, At2g26730, AGL42, JOINTLESS, AGAMOUS, MADS3, DEK44, PKSA, PKSB, CYP704B1, ABCG26, TKPR, and FAR2. These DEGs are mainly involved in the regulation of sugar synthesis, pollen fertility, seed development, and flowering stage; these genes were upregulated in the soybean triple mutants. Overall, our results identify several genes that can be modified to develop earlier-flowering and higher-yielding soybean cultivars.

Materials and methods

Phylogenetic tree analysis of the EIN2/EILs family

All soybean (cultivar Williams 82, W82) EIN2 and EIL (EIN3-like) protein sequences were downloaded from Phytozome13 (https://phytozome-next.jgi.doe.gov/) or the Plant Transcription Factor Database (http://planttfdb.gao-lab.org/) and aligned with those of A. thaliana, Cicer arietinum, Glycine soja, Lotus japonicus, Medicago truncatula, Oryza sativa, Phaseolus vulgaris, Prunus persica, Salix purpurea, and Vitis vinifera using the Muscle program within the MEGA 7.0 software package. A phylogenetic tree was constructed using the neighbor-joining (NJ) method with key parameters set as follows: P-distance model, pairwise deletion, and bootstrap = 1,000.

Plasmid construction

We designed the sgRNAs of EIN2 and EIL using the CRISPR direct online tool (http://crispr.dbcls.jp/), and their target sequences are listed in Table S7. Pairs of DNA oligonucleotides of the three sgRNAs were synthesized by the Shanghai Sangon Biological Engineering Technology and Service Company (Shanghai, China) and annealed to generate dimers, which were then integrated into Atu3b, Atu3d, Atu6-1, and Atu6-29 sgRNA expression cassettes by overlapping PCR. The target expression cassettes were successfully cloned into pYLCRISPR-Cas9Pubi-B using the Golden Gate cloning method [18]. Cas9 was expressed under the ubiquitin promoter, and positive plasmids were introduced into Agrobacterium tumefaciens strain EHA105 for stable transformation (Figure S2).

Stable soybean transformation

The soybean cultivar W82 was transformed using the Agrobacterium-mediated soybean cotyledon node [24]. T0 generation gene edited soybean plants were obtained following 2–3 months of regeneration and selected with 3–5 mg/L glufosinate-supplemented medium. Gene edited plants were confirmed (gene edited T0 and T1 plants) by spraying the total plants with 900 mg/L of glufosinate ammonium. Gene edited glufosinate-resistant (GR) soybean materials were then subject to amplicon deep sequencing of all genome-edited sites.

Amplicon deep sequencing of all genome-edited sites

The amplicon PCR products of T0 lines for next-generation sequencing were obtained through two steps. First, we performed 25-μL reactions each with soybean genomic DNA to amplify the sgRNA in the target regions. Second, Illumina adaptors (i7 and i5 primers) were attached and to barcode samples in a 25-μL reactions volume using 2 μL of the product from the first PCR step. The sequences obtained by multiplex amplicon sequencing in the single nucleotide polymorphism (SNP) genotyping are listed in Table S7. The synthesis of primers and amplicon deep SNP sequencing of all genome-edited sites were performed by Shanghai Sangon Biological Engineering Technology and Service Company (Shanghai, China). The raw sequencing data were filtered for quality control purposes with default parameters [49]. All merged reads were subsequently aligned to the reference using the BWA-MEM algorithm [50].

Phenotypic characterization of wild-type and the Z4 triple mutant plants

In the constitutive triple-response mutant lines for ethylene, we identified Cas9-free lines carrying targeted mutations, in which the T-DNA was lost while the EIN2L, EIL3, and EIL4 genes were transmitted. A pot experiment was conducted outdoors at Jilin Normal University, Siping City, Jilin Province, between 2019 and 2021 using the T2 generation seeds of the W82 and its EIN2L, EIL3, and EIL4 homozygous triple mutants (abbreviated as experiment number Z4). Plastic pots with an upper diameter, lower diameter, and height of 24, 17, and 20 cm, respectively, were used for cultivation. Twenty 1–5-cm-diameter holes were made at the bottom of the pots, and each pot was filled with approximately 6.0 kg of the cultivation medium (a 3:1 mixture of soil and vermiculite). The spacing between pots was uniformly and consistently 30 cm. To reduce border effects, potted W82 were placed around the experimental potted plants. W82 and Z4 plants were randomly arranged on the inner side of the border row. Three W82 or Z4 seeds were sown in each of the 30 pots (sowing depth = 3.0 cm) and one seedling was retained after germination. When the first flower and pod were observed on each plant, the corresponding flowering time and seed-setting stage were recorded for 10 individual plants, respectively. At the seed-maturity stage, physiological indexes such as plant height, above-ground dry weight, pod number, bean number, 100-seed weight, and yield per plant were measured for at least 10 individual plants. All the developmental periods were determined according to the literature [51].

Transcriptome sequencing of wild type and gene-edited soybean

Z4 and wild-type plants were grown under the conditions described for the phenotypic analyses above. Also, some of them were used to carry out RNA extraction and transcriptome sequencing. When the potted wild-type soybean flowered (approximately 35 d after cultivation), the terminal buds of the wild-type control and the Z4 plants were sampled. To investigate gene expression changes between the wild-type and gene-edited soybean, we constructed six DGE profiling libraries: CK-1, CK-2, CK-3, Z4-1, Z4-2, and Z4-3. Thus, our experiment included two treatments (WT and triple mutation), each treatment included three replicates, and each replicate included five terminal buds. Total RNA was extracted from the flower bud samples using the RNA EasySpin Isolation System (Aidlab Biotech, Beijing, China), and RNA pretreatment and mRNA sequencing were carried out as previously described [52]. Sequencing was performed using a HiSeq 4000 system (Illumina, San Diego, CA, USA), and raw transcriptome data were deposited in the Sequence Read Archive (https://www.ncbi.nlm.nih.gov/SRA) under BioProject number PRJNA774476.

Bioinformatics analysis of transcriptome data

High-quality clean data were obtained by filtering the raw data, which removed low-quality adapter sequences and reads. Cleaned data were then aligned to the soybean reference genome (https://soybase.org/data/public/Glycine_max/Wm82.gnm2.DTC4/) to generate the mapped data. ‘HISAT2’ (http://ccb.jhu.edu/software/hisat2/index.shtml) was used to map the RNA-seq reads [53], and ‘StringTie’ (https://ccb.jhu.edu/software/stringtie/index.shtml) was used to assemble the mapped reads [54]. Gene function was annotated based on Nr (NCBI non-redundant protein sequences), Nt (NCBI non-redundant nucleotide sequences), Pfam (Protein family), KOG/COG (Clusters of Orthologous Groups of proteins), Swiss-Prot (a manually annotated and reviewed protein sequence database), KO (KEGG Ortholog), and GO (Gene Ontology) databases. FPKM (fragments per kilobase of transcript per million fragments mapped) normalization was applied to measure the expression level of transcripts by StringTie using the maximum flow algorithm. DEGs were identified using ‘DESeq2’ [55]. The criteria for DEGs were set as FC ≥ 2 and FDR < 0.01. FC refers to the ratio of gene expression in the two samples; and FDR refers to the adjusted p-value, which was used to measure the significance of differences. Gene set enrichment analysis [56] was performed on all genes based on their expression levels, and the gene sets of KEGG pathways and GO terms for BP, CC, and MF were targeted as the gene sets of interest. Genes from each group were used as the background gene set, and enriched gene sets were identified where p < 0.001 and FDR < 0.05.

qRT-PCR

Using the same samples that were used for RNA-seq analysis, the RNA Easy Spin Isolation System (Aidlab Biotech, Beijing, China) was utilized to extract total RNA from the soybean plants according to the manufacturer’s instructions. In total, 12 DEGs and non-DEGs were chosen for qRT-PCR analysis, and primer designing and qRT-PCR analysis were carried out as previously described [57]. The detailed primer sequences are listed in Table S8. Beta actin (ACTB) was used as a reference gene in the qRT-PCR assay, and the 2^−ΔΔCt method was used to calculate the FCs of the chosen genes [58].

Statistical analysis

Analysis of variance (ANOVA) or generalized linear modeling (GLM) in SAS version 8.01 (SAS Institute, Inc., Cary, NC, USA) were used to test for differences in phenotypic data between multiple groups. The means of the data obtained from the control and gene-edited plants were compared using the least significant difference test at 5% significance level.

Supplementary Information

Below is the link to the electronic supplementary material.

Authors’ contributions

Yunqing Cheng, Jianfeng Liu, and Xiangdong Yang: Study conception, data collection, analysis and interpretation. Jianfeng Liu and Xiangdong Yang: Writing—original draft. Yujie Li, Hongli He, and Xingzheng Zhang: Preparing samples. All authors have read and approved the manuscript.

Funding

This work was financially supported by the National Natural Science Foundation of China (No. 32172079) and the Department of Science and Technology of Jilin Province (No. 20200201113JC).

Availability of data and materials

The original contributions presented in the study are publicly available. All raw transcriptome data were deposited in the sequence read archive (accession no. PRJNA774476).

Declarations

Ethics approval and consent to participate

This article does not contain any studies with human or animal subjects. The current experimental research and field study, including the collection of plant material, complies with relevant institutional, national, and international guidelines and legislation and is used for research and development. The use of all plant materials does not bring any risk to other species in nature.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interest.