Base Editing of EUI1 Improves the Elongation of the Uppermost Internode in Two-Line Male Sterile Rice Lines

Abstract

The use of male sterile lines (MSLs) of rice is essential for heterosis utilization. However, MSLs have a common defect in the elongation of the uppermost internode, eventually leading to incomplete panicle exsertion, blocking pollination, and reducing the hybrid rice seed yield. Previously, the elongated uppermost internode 1 (EUI1) was identified as an active gibberellin-deactivating enzyme that plays a key role in panicle exsertion from the flag leaf sheath in rice (Oryza sativa L.). We used an adenine base editor to edit EUI1 and obtained two types of homozygous transgenic plants (eui1-1 and eui1-2). The transcription and translation levels of EUI1 in the two mutants were significantly lower than in the wild type, as was the oxidation activity of EUI1 to active gibberellins (GAs), which also decreased. The contents of the plant hormones GA₁, GA₃, and GA₄ in eui1-1 (1.64, 1.55, and 0.92 ng/g) and eui1-2 (0.85, 0.64, and 0.65 ng/g) panicles were significantly higher than the wild type (0.70, 0.57, and 0.42 ng/g). The uppermost internode lengths of the mutant were 26.5 and 23.6 cm, which were significantly longer than that of the wild type (18.0 cm), and the cell lengths of the mutant were 161.10 and 157.19 μm, which were longer than that of the wild type (89.28 μm). Our results indicate that the adenine base editing system could increase the content of endogenous bioactive GAs in young panicles by fine-tuning EUI1 activity, reduce the defect of panicle enclosure in MSLs and increase the yield of hybrid rice seed production.

1. Introduction

It is well known that hybrid rice has a substantially increasing rice grain yield [1,2]. However, the male sterile lines (MSLs) of rice have a defect in the elongation of the uppermost internode, leading to incomplete panicle exsertion, which interferes with the pollination process and reduces the yield of hybrid rice seeds [3]. As a result, a large amount of exogenous GAs are required for hybrid rice seed production, increasing the cost and the probability of pre-harvest sprouting (PHS), and affecting the quality of hybrid seeds [4]. Therefore, it is critical to elucidate the molecular mechanisms to fine-tune panicle exsertion and provide substantial benefits for hybrid rice seed production.

Gibberellins (GAs) are a group of important diterpenoid compounds, some of which act as growth-promoting hormones that control diverse processes such as stem elongation, leaf expansion, seed germination, and flowering [5,6]. In rice, some mutants with altered GAs metabolism or signaling pathways have been studied [7,8]. GA biosynthesis involves three classes of enzymes: plastid-localized terpene cyclases, membrane-bound cytochrome P450 monooxygenases (P450s), and soluble 2-oxoglutarate-dependent dioxygenases (2ODDs). The bioactive GAs (GA₁ and GA₄) are deactivated by GA2 oxidases [9]. A prior study identified that EUI1 encoded a cytochrome P450 monooxygenase, CYP714D, which catalyzes the 16a,17-epoxidation of non-13-hydroxylated GAs and reduced the biological activity of GA₄ in rice [10].

Okuno and Kawai obtained the long internode recessive mutant, LM-1, by mutagenesis of Nong Lin 8 [11]. In 1981, a recessive rice mutant, eui1 (elongated uppermost internode1), was characterized by an extremely elongated uppermost internode and excess panicle exsertion at the heading stage [12]. Yang (2001) directly mutated the sterile (S), maintained (B), and restored (R) lines with Cobalt-60 (60Co) [13] and obtained a recessive mutant with internode elongation, and a non-allelic eui1 known as eui2 [14]. Zhu (2006) found an unrecognized pathway for GAs deactivation by EUI, and the eui mutant exhibited an extremely elongated uppermost internode [5]. Previous studies identified the strongest expression tissue of the EUI1 gene in young panicles, followed by the flag leaf and the uppermost internode, with a negative regulatory effect on the GA-mediated cell elongation of the uppermost internode in rice [5,10,15]. It was determined that when the EUI1 gene lost its function, the content of active GAs in young panicles significantly increased, and the active GAs produced in young panicles were transported downward, resulting in a highly significant elongation of the uppermost internode during the heading stage. Therefore, the eui1 mutation has been used to genetically improve the heading performance of MSLs [16,17,18]. In the conventional method, eui1 is used in cross- and backcrosses in sterile and restorer lines. For example, Yang (1998) proposed to mutate B, R, and S lines to obtain high recessive mutants and produce hybrid seeds [19]. However, due to the recombination of the gene, sterile and restorer lines lost some of their original advantageous agronomic traits, and excessive elongation of the uppermost internode resulted in lodging of the rice, increasing the risk of seed production. At present, the function of EUI1 on the elongation of the uppermost internode is clear, but the molecular mechanism of the moderate elongation of the uppermost internode by fine-tuning EUI1 through gene editing remains unclear [20,21,22,23].

Base editing is a new genome editing tool based on CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats). Compared with the CRISPR editing system, the base editing system can accurately edit the target gene. Currently, different plant base editors have been established by fusing various nucleobase deaminases with Cas9, Cas13, or Cas12a (Cpf1) proteins. The cytosine base editor (CBE) uses the action of cytosine deaminase to convert the amino group in the cytosine structure to an oxygen atom while losing the hydrogen atom in the adjacent nitrogen atom, so that cytosine (C) is converted to uracil (U), which is converted to thymine (T) in the process of DNA replication and repair [24]. The adenine base editor (ABE) uses adenine deaminase to convert the amino group in adenine nucleoside into an oxygen atom. At the same time, it loses the hydrogen atom on the adjacent oxygen atom to convert adenine (A) into hypoxanthine (I), which is then converted to guanine (G) in the process of DNA replication and repair, and finally achieves the conversion from adenine (A) to guanine (G) in the target region [25]. RNA base editors can also induce the base substitution of adenine (A) to inosine (I) or cytosine (C) to uracil (U) [26]. Base editing has been widely used in the animal, microbial, and gene therapy fields and has made great progress in crop breeding. At present, a single base editing system suitable for rice and wheat has been established and optimized to replace bases accurately and efficiently within the rice and wheat genomes. Zong (2017) used the CBE base editing system to perform single-base editing for breeding wheat, maize, and rice, obtaining improved plants with multiple favorable traits [27]. Li (2018) used an ABE base editing system to obtain point mutations by introducing functions, and an endogenous gene was successfully modified to improve herbicide resistance in rice [28].

In this study, we successfully mutated EUI1 by ABE, obtained two mutants eui1-1 and eui1-2 with elongated internodes in the uppermost internode, and reduced the panicle enclosure rate. Upon further investigation, we determined that the protein structure of EUI1 was changed in mutant eui1-1 and eui1-2, and that the translation levels decreased along with a decreased enzyme activity. We also detected a higher content of Gas and longer cells, which is consistent with the phenotype in eui1-1 and eui1-2. Our research demonstrates that the base editing mutation of EUI1 can fine-tune the expression of EUI1, weaken the oxidation ability of EUI1 to bioactive GAs, increase GA transport to internodes in young panicles, and improve the heading ability of sterile lines.

2. Materials and Methods

2.1. Plant Materials and Growth Conditions

The eui1-1 and eui1-2 mutants were generated from C815S thermosensitive genic male sterile (TGMS) lines. The T₁ or more advanced generated homozygotes from eui1-1 and eui1-2 mutants were used for evaluating the resulting characteristics. In this study, all plants were grown in an experimental field or greenhouse under natural conditions at the China National Rice Research Institute.

2.2. Vector Construction and Plant Transformation

To knock out the EUI1 gene via the adenine base editor (ABE), single-guide RNAs (sgRNA) targeting EUI1 (LOC_05g40384) genes were cloned into the ABE vectors, pH-PABE-7-esgRNA [28]. This CRISPR/Cas9-EUI1 construct was then transferred into C815S callus through an Agrobacterium tumefaciens-mediated transformation. Sequences of the primers used for vector construction and detection are listed in Supplementary Table S1.

2.3. RNA Extraction and qRT-PCR Analysis

To evaluate the relative gene expression levels in the transgenic plants, qRT-PCR was performed. Total RNA was isolated from roots, stems, leaves, panicles, and developing grains, 9, 12, and 15 days after fertilization (DAF) using Trizol reagent (Invitrogen, Carlsbad, CA, USA). The oligo (dT18) and the ReverTra Ace qPCR RT Kit (Toyobo, Osaka, Japan) were used for cDNA synthesis. The Ubiquitin gene was used as the internal control (GenBank accession No. AF184280), and a qRT-PCR experiment was performed with a LightCycler 480 system (Roche, Basel, Switzerland). Relative gene expression was calculated by the 2^−∆T method. The primer sequences used in this analysis were listed in Supplementary Table S1.

2.4. Protein Extraction and Immunoblotting

For extraction of total proteins, young panicles were harvested and homogenized in 1 mL of extraction medium. Extracts were centrifuged at 14,000 rpm for 15 min, and the supernatant was collected. The protein concentration of the supernatant was determined using a Pierce BCA protein assay kit (Thermo, Fisher, Waltham, MA, USA). Equal amounts of protein were loaded onto the SDS-PAGE gels, transferred to PVDF membranes, and immunoblotted with corresponding antibodies. The soluble and insoluble proteins were further separated from young panicles, as previously described [29]. Soluble and insoluble protein extracts were loaded onto SDS-PAGE gels with β-actin protein as a loading control, further transferred to PVDF membranes, and immunoblotted with the corresponding antibodies.

2.5. Light Microscopy

Stems and leaves were harvested from wild-type and transgenic mature plants. Stem tissues were treated with 15% of hydrofluoric acid before dehydration. The procedures of dehydration, clearing, infiltration, and embedding were carried out as mentioned above. The microtome sections (10 μm) were mounted on glass slides for imaging.

2.6. GA Bioactivity Assay

At the booting stage, young panicles of eui1-1, eui1-2, and wild type were collected, and three samples of each material were frozen in liquid nitrogen and stored at −80 °C for later use. The bioactivities of GA₁, GA₃, and GA₄ were determined.

2.7. CYP714D1 ELISA

Young rice panicles sampled at the booting stage were frozen and stored in liquid nitrogen. The EUI1 enzyme activity was determined using the plant cytochrome monooxygenase CYP714D1 enzyme-linked immunoassay kit (Cat. No.: WLB-3595301). The absorbance (OD) of the samples was measured at 450 nm using an enzyme-labeled instrument (Infinite^® 200 PRO NanoQuant), and the activity concentration of CYP714D1 in the samples was calculated according to the standard curve.

3. Results

3.1. EUI1 Base Editing

The base editor ABEs, used to perform the A g base substitution for the encoding region of EUI1 gene sequences, were determined via the rice genome annotation project (rice.plantbiology.msu.edu), and a 20-bp target sequence rich in adenine and guanine on the second exon of EUI1 was designed using the CRISPR/CAS9 base genome editing tool (skl.scau.edu.cn/targetdesign) (Figure 1A,B). The selected target sites were constructed into the vector (pH-PABE-7-esgRNA). The transgenic T₀ mutants were obtained after hygromycin and target site mutation detection, and the results show that EUI1 has two distinct types of mutations in the background of C815S, replacing one or two adenines with guanine (Figure 1B,C). Two types of homozygous mutants continue propagation to T₁ generation. According to the miss rate forecasting tool, select the three most likely miss genes (http://skl.scau.edu.cn/offtarget/) and the miss rate detection for the T₁ generation of mutants. The results showed that the 30 mutant plants did not have mutations in the other three genes (Supplementary Table S2) and that these plants were used for propagation of T₂ generation and related tests. The two mutation types changed the species and quantity of the original amino acid in EUI1 (Supplementary Figure S1). Furthermore, the protein spatial structure prediction website (sbg.bio-ic.ac.uk) demonstrated that the spatial structure of EUI1 changed in both mutant types (Figure 1E).

3.2. Panicle Enclosure Rate and Cell Elongation

Through agronomic trait observations of eui1-1 and eui1-2 T₁ plants, slight but significant differences in plant type between the mutants eui1-1, eui1-2, and the wild-type C815S were found (Figure 2A). However, the uppermost internode lengths of eui1-1 and eui1-2 were 26.5 and 23.6 cm and 8.5 and 5.6 cm longer than C815S, respectively. The elongation difference between other internodes was not significant. (Figure 2B,E). Although the uppermost internodes of the mutants were significantly elongated, the panicle wrapping rate was reduced, and the panicles were not fully extended (Figure 2C,D). Subsequently, we analyzed the cell length, width, and area in the longitudinal section of the uppermost internode of C815S and the two mutants. The results show that the cell length and area of eui1-1 and eui1-2 were significantly elongated. (Figure 2F–I). The elongation of cells at the uppermost internode increased the panicle exsertion rates of eui1-1 and eui1-2.

3.3. Transcription and Protein Content in eui1 Mutant Plants

A reverse transcription polymerase chain reaction (RT-PCR) was used to detect the transcription level of EUI1 in different rice tissues (roots, stems, leaves, sheaths, and young panicles) and seeds after 3, 5, 6, and 9 days (Figure 3A). EUI1 was expressed in different tissues, indicating that it is constitutively expressed and has the highest transcription level in young panicles. The RNA of eui1-1, eui1-2, and C815S young panicles was extracted, and a RT-PCR was performed to detect the transcription level of EUI1 genes in the young panicles of C815S, eui1-1, and eui1-2. Compared with the wild types (C815S), the expressions of EUI1 in eui1-1 and eui1-2 were reduced by 42.9% and 32.2%, respectively (Figure 3B). In general, the site-directed base substitution changed the expression levels of EUI1, thus validating the success of EUI1 gene editing. Western blotting was used to detect the contents of EUI1 in mutants and wild types, confirming the fact that the EUI1 content was reduced in both mutants (Figure 3C).

3.4. CYP714D1 Enzyme Activity and Active GAs Content

In the bioactive GAs biosynthetic and catabolic pathways, EUI1 inactivates GA₄, GA₉, and GA₁₂ through epoxidation (Figure 4A), thereby negatively regulating the elongation of the uppermost internode. To determine whether the base replacement affects the function of EUI1, we tested the CYP714D1 enzyme activity and the content of active GAs (GA₁, GA₃, and GA₄) in 815S, eui1-1, and eui1-2. The protein abundance and enzyme activity of EUI1 in eui1-1 and eui1-2 were significantly reduced (Figure 4B). The content of active GAs (GA₁, GA₃, and GA₄) in young panicles was detected via enzyme-linked immunoassays. Not only did the content of GA₄ significantly increase, but the content of GA₁ and GA₃ also increased (Figure 4C).

3.5. Analysis of Outcrossing Seed Setting Rate

In order to detect the outcrossing seed setting rate of eui1-1 and eui1-2, C815S, eui1-1, and eui1-2 were used as the female parents and R261 was used as the male parent for the hybrid seed production experiment (Figure 5A). Under the premise of the same seed production environment and without the application of exogenous gibberellin, statistics were conducted on the setting rates of C815S, eui1-1, and eui1-2, respectively. The results show that the setting rate of C815S (18.43%) was significantly lower than that of eui1-1 and eui1-2 (32.14% and 32.45%, respectively) (Figure 5B). Meanwhile, the panicle wrapping rate of C815S (17.91%) was significantly higher than that of eui1-1 (10.11%) and eui1-2 (12.80%) (Figure 5C). The above results show that the base-edited transgenic plants eui1-1 and eui1-2 showed a lower panicle wrapping rate and a higher seed setting rate without the application of exogenous gibberellin. The base editing of EUI1 can improve the hybrid seed production potential of MSLs.

3.6. Analysis of EUI1 in 3K

Given that EUI1 was likely selected during breeding, we then asked if they had been subjected to co-selection. To this end, we analyzed the allelic distribution of EUI1 in a panel of 3024 varieties reported in the 3K Rice Genomes Project [30,31]. For EUI1, there were six of that haplotype, but there are mainly two types of transcribed and translated proteins. Hap1 was the most common among indica and japonica varieties, and the EUI1 genotype showed no specific differentiation pattern between indica and japonica, which proved the genetic stability of EUI1, and except Hap1, a few Hap2, Hap3, Hap4, Hap5, and Hap6 genotypes may be the result of artificial and natural selection in indica and japonica rice (Figure 6A,B). This indicates that it is feasible and has a broad adaptability to editing the EUI1 gene and improving the panicle elongation of MSLs by base editing.

4. Discussion

MSLs have aggressively been used to produce hybrid rice seeds for increased rice yield. Though MSL cultivars have been used in producing hybrid seeds, this hinders internode growth and results in panicle enclosure within leaf sheaths at the heading stage. To counter this, exogenous GA₃ treatments are used to stimulate panicle emergence in MSL cultivars. However, this stimulates the preharvest sprouting of grains and shortens the longevity of hybrid seeds due to GA₃ activity. Therefore, breeding genetic materials that can improve panicle extension from the leaf sheath of MLSs is key to increasing the yield of hybrid seeds. According to the literature, some genes play a role in elongating the uppermost internode, including EUI1, which can inactivate GA₄ through epoxidation and negatively regulate the elongation of the internode [22,32]. The eui1 germplasm was used to improve the ability of the panicle extension of sterile lines [3,17,18]. However, the excessive elongation of the uppermost internode increases the risk of lodging, making this application impractical for hybrid seed production (Supplementary Figure S2).

The direct mutagenesis of EUI1 using physical or chemical mutagens will drastically affect EUI1 function, preventing the precise regulation of the elongation of the uppermost internode. Therefore, fine-tuning EUI1 can better adjust the elongation of the uppermost internode. The replacement of one or more bases does not destroy the integrity of the protein. By changing the protein structure of EUI1, it is possible to fine-tune the oxidative activity of active GAs, thereby moderately increasing their content and somewhat extending the uppermost internodes. In this study, we used a new gene editing technology (ABE) to perform gene editing on the CDS sequence of the EUI1 gene through base substitution and obtained two transgenic materials with different base substitution types, eui1-1 and eui1-2 [33]. The mutation type of eui1-1 had only one amino acid change, while eui1-2 had two amino acid changes along with a change in the structure of the EUI1 protein. The plant height and plant type of the mutants, eui1-1, and eui1-2 had no significant difference compared with the wild type, but the uppermost internodes were significantly elongated and the leaf sheath rate was significantly reduced.

To investigate whether the differences in the leaf sheath were influenced by the base editing of EUI1 in transgenic plants, we explored the EUI1 gene expression levels in C815S, eui1-1, and eui1-2, revealing that EUI1 was significantly reduced in the two mutants. We further investigated the protein content of EUI1 and found it was slightly reduced in eui1-1 and eui1-2, demonstrating the effect of base editing on reducing the transcription and translation levels of EUI1.

Base substitution changed the type of amino acids, and the EUI1 protein structure changed after the mutation, as determined via a protein structure prediction. The enzyme activity of EUI1 in eui1-1 and eui1-2 was tested, showing that the EUI1 enzyme activity decreased in the mutants. Considering the increase in active GAs in eui1-1 and eui1-2 mutants and the elongation phenotype of the uppermost internode, we conclude that base substitutions reduced the transcription, translation, and activity of EUI1. Furthermore, decreased EUI1 protein content and enzyme activity weakened the inactivation effect of EUI1 on active GAs, leading to the accumulation of active Gas and in the elongation of the uppermost internodes.

Although the uppermost internodes of eui1-1 and eui1-2 mutants were significantly elongated, the wrapping rate of panicles decreased, and the panicles were not fully extended. While the upper internode stretches, the leaf sheath also stretches to a certain extent, inhibiting full panicle extension. Thus, it is essential to understand how to regulate the elongation of the uppermost internode while restricting the elongation of the leaf sheath. In this study, we found that one or two amino acid substitutions did not significantly change EUI1 protein activity, and we hypothesize that base substitution at additional sites can result in panicle extension.