穗部小花结实率是影响谷类作物产量的关键因素,相关的调控机制至今仍不清楚。本研究鉴定到一个谷子花器官败育无籽突变体（sinog1），解剖学分析表明sinog1突变体的穗部小花在抽穗期发生褐变并逐渐败育。通过构建BC₃F₂群体，将sinog1的候选基因定位在5号染色体32.44Mb至34.29Mb的1.89Mb区间；RNA-seq分析表明野生型和突变体的差异表达基因主要富集在ABC转运体通路，结合BSA-seq和RNA-seq分析发现候选区间内的3个ABC转运体通路相关候选基因在差异表达基因中富集，推测ABC转运体通路可能是影响sinog1表型的关键通路。本研究为深入了解谷子以及近缘禾谷类作物的小花发育机制奠定了理论基础。

Increasing crop grain yields is an urgent global priority due to population growth, shrinking arable land, and severe climate change in recent years (Tang et al. 2023).  Unraveling the process of panicle development is crucial for enhancing the grain yield of cereal crops.  In the development of rice panicles, the inflorescence meristem (IM) gives rise to two types of lateral branch meristems (BMs): primary branch meristem (pBM) and secondary branch meristem (sBM).  The pBM generates sBM and spikelet meristems (SMs), and the sBM further differentiates into more SMs (Zhang and Yuan 2014).  A spikelet is the basic unit of inflorescence in Poaceae plants.  It originates from the spikelet meristem (SM) that determines the number of spikelets per panicle and consequently impacts crop yield formation (Doebley et al. 2006).  The seed setting rate, determined by spikelet development, is another crucial trait intimately linked to grain yield (Zhuang et al. 2024).  In plants, developmental defects in spikelets are typically categorized as either male sterility (Notsu et al. 2002; Liu et al. 2007; Luo et al. 2013) or female sterility (Li et al. 2022).  However, no mutant with completely sterile male and female reproductive organs has yet been identified so far.

Foxtail millet (Setaria italica (L.) Beauv.) is a stress-tolerant annual cereal crop species from the Poaceae family (Muthamilarasan and Prasad 2015; He et al. 2023; Liang et al. 2023).  In recent decades, foxtail millet has been developed as a new model crop for deciphering panicle development due to its diploidy and small genome, short growth cycle, and self-pollination character (Doust et al. 2009; Diao et al. 2017; He et al. 2021).  In foxtail millet, each spikelet produces one fertile floret and one sterile floret.  The sterile one was degenerated, and the fertile one was enclosed by lemma, palea, and two lodicules derived from the sterile one (Hussin et al. 2021; Zhang et al. 2021).  Several abnormal panicle mutants have been identified in both foxtail millet and its wild type, green foxtail (Setaria viridis).  In foxtail millet, the silp1 mutant showed an increase in the length and width of primary branches, in company with a decrease in the number of fertilized spikelets and seed setting rate in the panicle (Xiang et al. 2017).  The siaux1-1 mutant displays a sparsely branched panicle compared to the wild type (Tang et al. 2021).  The simads34 mutant shows increased panicle width and decreased panicle length and grain yield in foxtail millet (Hussin et al. 2021).  In green foxtail, mutations in the svaux1 gene result in reduced inflorescence branches and spikelet numbers and increased panicle length (Huang et al. 2017).  Approximately 17% of the panicles in the brl1 mutant produce additional flowers, bristles and/or spikelets within each spikelet compared to one flower per spikelet in the wild type (Yang et al. 2021).  The inflorescences of the Svfon2 mutant exhibit abnormal apices and panicle tips, which are divided into two or more parts (Zhu et al. 2021).  However, few studies on panicle sterility mutants have been reported in both Setaria italica and Setaria viridis, with only one male-sterile mutant, sinp1, having been identified (Zhang et al. 2021).  Thus, the understanding of panicle infertility in foxtail millet remains limited.

In this trial, we identified a completely sterile mutant sinog1 (no grain 1) from the EMS-induced mutant library of Yugu 1.  The mutant exhibited a slender panicle and was completely sterile with no seed set (Fig. 1-A).  Dissection observations revealed that all the reproductive organs of sinog1 florets gradually turned brown and eventually failed at the heading stage, leaving only two lemmas in the florets of sinog1, while the wild type was able to form mature floral organs normally (Fig. 1-B).  Additionally, the spikelet of the sinog1 mutant was significantly narrower and longer compared to the wild type (Fig. 1-A and C).  Furthermore, the spikelet of sinog1 mutant exhibited a significant increase in the content of indoleacetic acid (IAA) ((44.20±0.96) ng g^–1 FW), abscisic acid (ABA) ((91.27±1.77) ng g^–1 FW), and gibberellic acid 4 (GA₄) ((2.04±0.03) ng g^–1 FW) compared to the wild type ((30.64±0.59) ng g^–1 FW, (63.96±1.53) ng g^–1 FW, and (1.88±0.07) ng g^–1 FW), and a significant decrease in the content of brassinosteroid (BR) ((1.85±0.05) ng g^–1 FW) compared to the wild type ((2.02±0.05) ng g^–1 FW) (Fig. 1-D).

To identify the candidate genes responsible for the complete sterility in sinog1, we constructed an F₂ population with 263A as the female parent and the residual heterozygote individuals, including the mutation of sinog1 (Deng3-3) as the male parent.  The residual heterozygote individual Deng3-3 was derived from the BC₃ population generated by multiple rounds of backcrosses with Yugu 1.  Genetic and segregation ratio analysis suggested that the complete sterile phenotype of sinog1 was caused by a single recessive gene (χ²=0.06<χ² (0.05,1)=3.84).  Combining BSA-seq (45× coverage) and linkage verification analysis of 343 recessive individuals (completely sterile plants), the causal gene of sinog1 was finally mapped to a 1.85 Mb interval from 32.44 to 34.29 Mb on chromosome 5.  A total of 203 genes with sequence variations were screened out within the 1.85 Mb genomic interval (Fig. 1-E).

Subsequently, the transcriptomes of the wild type and mutant panicle at the heading stage were analyzed, identifying 37,986 genes.  Among these genes, 5,014 are differentially expressed genes (DEGs) between the wild type and sinog1 (log₂FC≥1 and FDR≤0.01), consisting of 2,707 significantly up-regulated and 2,307 significantly down-regulated genes.  The KEGG pathways analysis revealed that pathways related to ABC transporters, photosynthesis, and diterpenoid biosynthesis were significantly enriched in DEGs (Fig. 1-F).  Among these pathways, ABC transporters have been confirmed to regulate floral organ formation and panicle growth by transporting phytohormones and heavy metals in plants (Do et al. 2018; Naaz et al. 2023).  For instance, the mutation of ABCG26 in Arabidopsis leads to severely reduced fertility (Choi et al. 2011).  In rice, the knock-down of OsABCB14 can decrease the concentration and polar transport rates of auxin, and iron concentrations are also increased in the mutant (Xu et al. 2014).  These results suggest that the mutation gene of sinog1 might be an essential regulator for the proper expression of ABC transporters involved in floral development in foxtail millet.

Among the 203 genes located in the candidate interval on chromosome 5, we identified 28 DEGs by RNA-seq (Fig. 1-G), including seven genes (Seita.5G274600, Seita.5G274500, Seita.5G270700, Seita.5G274000, Seita.5G283000, Seita.5G282700, and Seita.5G272800) that were highly expressed in the wild type and 21 genes that were highly expressed in sinog1.  Of the 28 common genes, ten had mutations in the exon regions, including Seita.5G283500 (Val₆₄₉Leu), Seita.5G283600 (Leu₂₄₇Ser, Asn₃₉₆Glu, and Ser₆₆₇Arg), and Seita.5G282700 (insert 6 bp, Ala₇₈Ser), which have been annotated as steroid/xenobiotic-transporting ATPase in foxtail millet and ABC transporter in rice and Arabidopsis thaliana.  Interestingly, these three ABC pathway-related candidate genes were also enriched in the KEGG pathway of DEGs in sinog1.  It is noteworthy that ABC transporters have been verified to contribute to auxin and ABA transport, response to heavy metals such as iron and aluminum, and regulation of stomatal characters, which are important for plant growth and development (Do et al. 2018; Naaz et al. 2023).  Therefore, we inferred that ABC transporters may be key genes influencing morphological mutations in sinog1.

In conclusion, we have identified a novel panicle sterile phenotype in foxtail millet, the sinog1 mutant, which exhibits complete sterility with aborted reproductive organs.  The morphological analysis confirmed that sinog1 showed complete flower abortion at the heading stage, and the candidate pathways contributing to this completely sterile phenotype were identified using the combined BSA-seq and RNA-seq methods in foxtail millet.  This research provides a theoretical basis for understanding florets development in foxtail millet and other crops.