Background

Plant inflorescence configuration is a key agronomic factor affecting mine grain yield and a major target for crop domestication and improvement [1]. Understanding the genetic basis of crop inflorescence structure will not only help to clarify the evolutionary mechanism of crops but also help to improve crop yield.

Rice (Oryza sativa. L) is one of the most important food crops in the world, and its panicle morphological development and molecular regulation mechanism have been the focus of research. In recent years, some progresses have been made in the study of panicle development of rice, a monocot model plant, but it is far less detailed than that Arabidopsis thaliana [2]. The panicle formation process of rice is a complex physiological and biochemical process involving axillary meristem development, inflorescence structure building and grain development. The in-depth study of panicle formation will not only help to reveal the regulatory mechanism of panicle morphogenesis but also provide theoretical guidance for the improvement of panicle type in rice [3]. Genes related to rice panicles, such as Gn1a/OsCKX2 [4], DEP1 [24], spr4 [25], spr5 [26], spr8 [27] and OsLG1 [1]. However, of these localized genes or QTLs, only OsLG1 was successfully cloned as the SPR3 locus [24]. Further studies demonstrated that a single nucleotide polymorpho-6 (SNP6) in the 11 kb cis-regulatory region upstream of the transcriptional start site of OsLG1 gene resulted in a compact panicle type in cultivated rice during rice domestication, but it did not change its expression in the leaf tongue, resulting in a compact panicle type and normal leaf tongue development in cultivated rice [1, Development of molecular markers

Insertion-deletion (InDel) markers were designed by manual comparison of genome sequences between japonica (cv. Nipponpare) and indica (cv. 93 − 11), and primers were designed to map the polymorphic regions of rice subspecies using Primer Premier 5.0.

PCR amplification and molecular marker detection

Plant DNA and DNA amplification was performed by polymerase chain reaction (PCR) with minor modifications [36]. The PCR products were separated by 8% polyacrylamide denaturing gel electrophoresis, and the molecular markers were stained with silver [37].

Bulked segregant analysis

Markers of target genes were identified by bulked segregant analysis. Leaf DNA of 20 mutant plants randomly selected from the F2 population were used for construction of a mutant DNA library. SSR markers distributed in the rice genome were used to amplify the spr9 mutant DNA, and Hui1586 DNA was used as a control for linkage detection. The marker band of the mutant gene was the same as the marker band of the spr9 mutant.

Molecular map** of the spr9 gene

The band types of the mutant (spr9 spr9) and wild type (SPR9 SPR9) were denoted as 1 and 3, respectively; 2 was used to represent a heterozygote (spr9 SPR9). Linkage analysis was performed using MAPMAKER version 3.0 software [38], and the linkage map was basically the same as the linkage map reported previously [39].

First, 326 SSR markers were screened from the rice molecular map to study the polymorphisms of spr9 and Hui1586. The results indicated that 253 pairs of primers showed polymorphism. Twenty plants with mutant phenotype and 20 plants like wild type selected from the F2 population, respectively, and linkage analysis of the spr9 locus was performed using these 253 polymorphic markers. Second, to narrow the map** region, we identified 1452 mutants from the F2 population of spr9 × Hui1586. By comparing Nipponbare and the indica cultivar 93 − 11 (http://www.elabcaas.cn/rice/index.html), InDel markers in the open rice genome sequence were designed to predict the likelihood of polymorphism between spr9 and Hui1586.

Bioinformatics analysis

Candidate genes were predicted based on existing sequence annotation databases (http://rice.plantbiology.msu.edu/; http://www.tigr.org/). Clones were fixed on the target gene combination mark with the basic local alignment search tool (BLAST) in pairs (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastn&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome) for sequence alignment. The DNA sequences of spr9 and SPR9 were used for a complete alignment with Clustal X version 1.81.

CRISPR mediated editing

The CRISPR-plant database and website were used to design gRNA spacer sequences with high specificity (Supplementary Table 1) [40]. A gRNA interval spanning the first exon of the gene was used to target the SPR9 gene of Hui1586. The transformation and identification of the edited rice lines were performed by Wuhan Boyuan Technology Company. After obtaining the edited transgenic plants, PCR products of transgenic CRISPR-edited lines were sequenced to identify specific mutations [41]. Primers for the CRISPR/Cas9 study are listed in Supplementary Table 1.

GFP fusion and subcellular localization

A GFP-SPR9 fusion was constructed 35 S: SPR9-pSuper1300-GFP and transformed to Agrobacterium tumefaciens GV3101, which was used to infiltrate tobacco leaves, and the expression of GFP was observed by laser confocal microscopy (Zeiss 880) .

Results

Analysis of the main agronomic traits of spr9

To elucidate the regulatory genes involved in spike development in rice, we screened and obtained a spr9 mutant with changes in spike traits in the R20-1 background, which displayed a spread panicle (Fig. 1a) and the corresponding wild type gene was named Spreading Panicle 9 (SPR9). Phenotype comparisons between the spr9 mutant and wild-type R20-1 are shown in Table 1. The results showed that there were no significant differences in plant height, panicle length, effective panicle number, number of grains per panicle, seed setting rate, 1000-grain weight, grain length or grain width between the spr9 mutant and the wild type (Fig. 1b-I and Supplementary Table 2).

Fig. 1
figure 1

Phenotypic comparison of the spr9 mutant and wild type R20-1. a: The spr9 mutant showed the phenotype of spreading panicles compared with R20-1. b, c, d, e, f, g, h, and i indicate no differences in plant height, panicle length, number of effective panicles, spikelets per panicle, seed setting rate, 1,000-grain weight, grain length and grain width between R20-1 and the spr9 mutant, respectively

Full size image
Table 1 Segregations of spreading panicle phenotype of the F2 populations produced from crossing of the spr9 mutant and R20-1
Full size table

Resistance analysis of RFS between the spr9 mutant and R20-1

A certain relationship was proposed between rice panicle type and the resistance level of RFS [30, 31]. To further compare whether there is a difference in resistance in RFS between the spr9 mutant and R20-1, we inoculated the spr9 mutant and R20-1 plants by manual injection, and each material was set up in triplicate. The experimental results showed that the disease score of the spr9 mutant was 3 and that of R20-1 was 4, further suggesting that the spr9 mutant exhibited more resistance to rice smut than the R20-1 control (Fig. 2a and b).

Fig. 2
figure 2

Disease symptoms of the spr9 mutant and R20-1 infected by U. virens. a: Rice smut resistance of the spr9 mutant and R20-1 plants using injected inoculation in a greenhouse. Three replications were carried out, each individual was artificially inoculated with U. virens in three panicles. Inoculated plants were kept in a greenhouse at 27 °C with 90–100% RH for 7 days. Then, they were placed at 27 °C and 80% RH until rice false smut symptoms appeared. b: The disease score of the spr9 mutant is 3 and that of R20-1 is 4. The spr9 mutant exhibited more resistance to rice smut than R20-1. Asterisks indicate statistical significance (p < 0.01) determined by Student’s t-test

Full size image

Genetic analysis of the spr9 mutant

To determine the genetic mode of spr9 mutant, we generated two kinds of hybrid F1 plants by crossing spr9 mutant with R20-1 and R20-1 with spr9, respectively. Both kinds of F1 plants showed same panicle phenotype as the wild type, indicating that the spr9 was recessive heredity and independent of cytoplasm (Table 1). Then in the two F2 populations, we found that the panicle phenotype of plants as spr9 mutant and R20-1 wild type fits a 1:3 segregation ratio (χ2 = 0.118～0.386, P > 0.5), implying a single gene controlled the spr9 mutant phenotype. Together, these results indicate that the spr9 mutant is genetically controlled by a single recessive gene.

Preliminary molecular map** of the spr9 gene

To determine the gene underlying the spr9 mutant, we further conducted a gene map** population by crossing the spr9 mutant with the japonica rice cultivar Hui1586. Total of 253 SSR markers showed polymorphism between spr9 and Hui1586 were obtained. Among these polymorphic markers, RM8211, on chromosome 5, showed a complete co-segregation with mutant phenotype in the selected 20 F2 plants with mutant phenotype and 20 plants with wild type phenotype. Furthermore, 193 recessive plants from the F2 population were genotyped by RM8211 and RM5970, recombinant plants were found, verifying the linkage relationship between the marker and spr9 mutant phenotype. Subsequently, we found one other polymorphic marker RM5970, located on chromosome 5, also showed linkage with spr9 mutant phenotype. The recombinants identified from RM8211 all showed homozygous genotype as spr9 mutant at the RM5970 locus, and the recombinants from RM5970 presented heterozygous or wild homozygous genotype at RM8211 locus, therefore, we preliminarily mapped the spr9 gene into the region of RM8211 and RM5970. Thus, spr9 was preliminarily located in a 16.5 cM region between SSR markers RM8211 and RM5970 on chromosome 5 (Fig. 3a).

Fig. 3
figure 3

Physical maps and structural comparison of the spr9 gene. a: Primary map** of the spr9 gene. The gene was mapped to the region between the markers RM8211 and RM5970. b: Further map** of the spr9 gene. The gene was mapped to the region between commonmarkers RM1090 and RM1187. c: Accurate map** of the spr9 gene. The spr9 was mapped to the region between markers Indel15-5 and Indel15-6 selected from 10 newly developed InDel markers. d: Fine map** of the spr9 gene. The spr9 was localized to a 43-kb region between the markers Indel15-18 and Indel15-22 selected from 21 newly developed InDel markers, and the recombinant number between the markers and target genes is indicated under the linkage map. e: Candidate genes in the 43-kb target region. f: SPR9 has three exons, and spr9 exhibits a 1 bp deletion in the first exon

Full size image

Fine map** of the spr9 gene

To localize the spr9 gene in a smaller region, we constructed a genetic map between RM8211 and RM5970. Two pairs of polymorphic primers RM1090 and RM1187 were screened from the common primers between RM8211 and RM5970. Furthermore, 193 recessive plants from the F2 population were genotyped by RM1090 and RM1187, recombinant plants were found, verifying the linkage relationship between the marker and spr9 mutant phenotype, spr9 was located between the RM1090 and RM1187 molecular markers, and the distance between the two molecular markers was 5.9 cM (Fig. 3b). To further localize the spr9 gene, 1452 recessive plants from the F2 population were genotyped by Indel5-3, Indel5-5, Indel5-6 and Indel5-10 (Table 2), recombinant plants were found, verifying the linkage relationship between the marker and spr9 mutant phenotype. The localization results showed that the spr9 gene was located between the molecular markers Indels 5–5 and 5–6, and the physical distance between the two markers was 477 kb (Fig. 3c; Table 2). To accurately locate the spr9 gene, 1452 recessive plants from the F2 population were also genotyped by 10 polymorphic InDel markers Indel5-12, Indel5-15, Indel5-16, Indel5-18, Indel5-20, Indel5-22, Indel5-25, Indel5-26, Indel5-29 and Indel5-31 (Table 2), recombinant plants were found, verifying the linkage relationship between the marker and spr9 mutant phenotype. The result showed 10 polymorphic InDel markers were detected 20, 9, 3, 1, 0, 1, 4, 8, 12 and 35 recombinant plants, respectively (Fig. 3d). Therefore, we accurately located the spr9 gene between the Indel5-18 and Indel5-22 molecular markers, and the physical distance between them was 43-kb (Fig. 3d).

Table 2 InDel and SSR molecular markers used for fine map** of the spr9 gene
Full size table

Candidate gene analysis in the 43-kb region

The functions of six candidate genes were annotated (LOC_Os05g38500, LOC_Os05g38510, LOC_Os05g38520, LOC_Os05g38530, LOC_Os05g38540 and LOC_Os05g38550) in this 43-kb region (Fig. 3e). According to the database annotation, the results showed that each of the six candidate genes had a corresponding full-length cDNA.

To find which gene is responsible for the mutant phenotype, we sequenced the above six genes in R20-1 and the spr9 mutant, and the results showed that there was only a 1-bp deletion (T) (LOC_Os05g38520) between wild-type R20-1 and the spr9 mutant (Fig. 3f). No difference were observed in the sequences of other five genes. Therefore, we hypothesized that LOC_Os05g38520 corresponds to SPR9. Open reading fragment analysis showed that the SPR9 gene (LOC_Os05g38520) had three exons and two introns (Fig. 3f).

The spr9 gene is responsible for the spread panicle phenotype

To determine the phenotype of spr9 in the japonica genetic background, we examined whether knockout of SPR9 in the cultivar Hui1586 (japonica) would lead to the spread panicle phenotype. Using the CRISPR/Cas9 gene editing system, a sequence-specific guide RNA (sgRNA) was designed to knock out the SPR9 gene. We obtained a total of three homozygotes from three independent knockout events and confirmed their presence of insertion and deletion mutations at the target sites by Sanger DNA sequencing (Fig. 4a). We then investigated and analysed the panicle characteristics of three homozygous lines after maturity and found that all three homozygous lines showed a panicle spread phenotype (Fig. 4b), which indicated that knockout of the SPR9 gene in Hui1586 would lead to the spread panicle phenotype. In addition, analysis of other agronomic traits showed that there were no significant differences in plant height, panicle length, effective panicle number, number of grains per panicle, seed setting rate, 1000-grain weight, grain length or grain width between Hui1586 and three knockout transgenic lines (Supplementary Table 3). These results were consistent with the agronomic traits of the spr9 mutant in the R20-1(wild-type). Taking together, we concluded that spr9 gene was the causal gene for the spread panicle phenotype in the spr9 mutant.

Fig. 4
figure 4

Knockout transgenic lines showed the phenotype of the spr9 mutant. a: Three independent events (designated SPR9-KO-Line1, SPR9-KO-Line2 and SPR9-KO-Line3) were generated using the CRISPR/Cas9 system and verified by sequencing. b: Panicle differences between Hui1586 and three knockout lines. The three knockout lines generated by CRISPR/Cas9 all exhibit the phenotype of spreading panicles

Full size image

Expression pattern and subcellular localizationof SPR9

To further understand the function of SPR9, reverse transcription- quantitative PCR (RT‒qPCR) was used to detect the expression patterns of SPR9 at different developmental stages of rice (Primers are shown in Supplementary Table 1). The results showed that SPR9 was expressed in all tissues tested here, including roots, shoots, and leaves of two- and four-week-old panicles of 0.5–1 cm, 1–3 cm, 3–5 cm, and 5–10 cm length, along with germinating and mature seeds and callus, but it was predominantly expressed in the seed (germination) and in panicles (5–10 cm length) (Fig. 5).

Fig. 5
figure 5

The expression patterns of SPR9. The expression patterns of SPR9. RNA samples were extracted from different tissues of Jiafuzhan, including roots, shoots, and leaves of two-, four- and six-week-old seedlings, spikelets of 0.5–1 cm, 1–3 cm, 3–5 cm, and 5–10 cm in length, germinating and mature seeds and callus. Data represent the mean and standard deviation of three biological replicates. Three technical replicates for each biological sample were used. The error bar represents the standard deviation (SD) of the value from three independent biological samples

Full size image

In order to further analyze the localization of SPR9, we constructed 35 S: SPR9-pSuper1300-GFP vector (Primers are shown in Supplementary Table 1) and transformed to Agrobacterium tumefaciens GV3101. Three days after injecting tobacco leaves, the localization of SPR9 cells was observed by laser confocal microscopy (The Zeiss 880 confocal microscope). The results showed the localization of SPR9 in the nucleus (Fig. 6).

Fig. 6
figure 6

The subcellular localizationof SPR9. The localization of SPR9-pSuper1300-GFP in Nicotiana tabacum L. cells was observed by laser confocal microscopy. The results showed that SPR9-pSuper1300-GFP mainly expressed in the nucleus, Bar = 20 μm

Full size image

Discussion

SPR9 is a new spread panicle-related gene

The spread panicle trait is a typical mutant trait in rice, and there have been some reports on its gene map**. To date, a total of 7 genes for spike traits have been reported, namely, spr1 [22], spr2 [23], spr3[24], spr4 [25], spr5[26], spr8 [27] and OsLG1[1]. Among these genes, spr1 and spr8 were controlled by a recessive nuclear gene, and other five sprouting trait genes were controlled by dominant genes. In this study, the spr9 gene was identified as a recessive mutation. Through further comparison, no cloned spread panicle gene was found in this region, so we speculated that spr9 was a new gene.

Why does spr9 show enhanced resistance to RFS?

SPR9 encodes the 60 S ribosomal protein L36-2, which is a ribosome associated protein. However, why does spr9 affect the disease resistance of rice? Studies have shown that posttranslational reprogramming is another fundamental regulatory pathway of plant immunity and specifically regulates translation in response to PAMP-triggered immunity induction [42]. Consistent with this result, previous transcriptomic studies of rice Pi21-silenced plants infected by Magnaporthe oryzae suggested that ribosomes were a third enrichment pathway compared to Nipponbare plants [43]. A recent study showed that translation was the most significantly enriched term in Gene Ontology (GO) analysis, while ribosomes were the most significantly enriched pathway in the Kyoto Encyclopaedia of Genes and Genomes (KEGG) analysis [46]. In rice, it was found that the spike characters were related to the occurrence of RFS. The rice varieties with vertical close ear type are more susceptible to RFS, while the rice varieties with scattered ear type and long curved ear type are not [30, 31]. We speculated that the main reason for the relatively better resistance of spread or long-curved rice panicle types is that under the same environmental conditions, spread or long-curved rice panicle types have better permeability, short duration of high humidity conditions, and relatively low humidity. More importantly, studies have shown that the best condition for rice false smut is low temperature and high humidity [47, 48]. Therefore, this is consistent with the results of this study that the spr9 mutants exhibit better disease resistance to RFS.

The application prospect of the spr9 gene in rice breeding

Many wild rice species have typical spread spikelets because the spread spikelets help them to improve the outcrossing seed setting rate and reproductive ability in the field to better adapt to the environment [49, 50]. At the same time, some studies have shown that the spread of panicle traits is often linked with some undesirable agronomic traits, such as shorter plant height, fewer tillers, fewer branches and stalks, lower yield, and stronger grain setting, but the grain quality is better[49, 50].

In this study, most importantly, our data indicated that the spr9 mutant not only enhanced resistance to RFS but also did not affect the important agronomic traits of rice (Figs. 1 and 2 and Supplementary Table 2). Together, the results indicate that SPR9 has good application prospects in future rice disease resistance and improved panicle breeding. For example, to improve the outcrossing rate of male sterile lines, we can transfer spr9 into sterile lines. At the same time, to improve rice smut resistance to a certain extent, we transferred spr9 into restorer lines.

Conclusion

In this study, a novel SPR9 gene was mapped and identified as a ribosomal protein coding gene. Then, CRISPR/Cas9 knockout experiments confirmed that the SPR9 gene is responsible for the spreading panicle phenotype of the spr9 mutant. Importantly, the spr9 mutant was found to improve resistance to RFS without affecting major agronomic traits, indicative of potential applications of spr9 in broader breeding programs. Taken together, our results revealed that the spr9 gene has good application prospects in future rice disease resistance and improved panicle breeding.