Genome-Wide Characterization of the PaO Gene Family and Pyramiding Effects of Superior Haplotypes on Yield-Related Traits in Sorghum

Abstract

The Pheophorbide a oxygenase (PaO) is a key enzyme in chlorophyll degradation and plays an important role in plant senescence. However, the PaO gene’s function in sorghum remains underexplored. In this study, we identified five SbPaO gene family members in the sorghum genome through bioinformatics analysis. Analyses of gene structure, phylogeny, and collinearity revealed high conservation of this gene family among grass crops, suggesting similar functions. Subcellular localization and protein network predictions indicated that SbPaOs may participate in chlorophyll catabolism and regulate leaf senescence. Expression pattern analysis showed that SbPaO1, SbPaO3, SbPaO4, and SbPaO5 were highly expressed in leaves and significantly upregulated during senescence. Haplotype analysis found three SbPaO genes significantly linked to thousand-grain weight (TGW); superior haplotypes SbPaO1-hap4, SbPaO3-hap5, and SbPaO4-hap4 notably increased this trait. Single-gene improvements increased TGW by 10.57–17.20%, dual-gene aggregation by 18.78–24.75%, and three-gene aggregation by 29.09%. The study also developed Kompetitive Allele-Specific PCR (KASP) markers that identify superior haplotypes with 100% accuracy. In summary, this study’s results provide a theoretical basis and genetic resources for further exploration of haplotype pyramiding strategies to breed new high-yielding sorghum varieties and delineate a clear research direction for subsequent functional validation and breeding practices.

1. Introduction

Sorghum (Sorghum bicolor L.) ranks among the top five food crops in the world. It is a staple food source for more than 500 million people worldwide. Sorghum has been recognized as a key crop for combating climate change and ensuring food security because it shows high tolerance to heat and drought stress [1]. In recent years, the use of sorghum has expanded into various fields. These include feed, brewing, forage, and biomass energy. As a result, it has become increasingly important in global agriculture [2]. Crop yield is governed by a number of factors. Among these, the senescence process has a significant impact on yield formation [3]. Senescence is a process of functional decline that occurs at the cellular, tissue, organ, or holistic level. It is accompanied by the translocation and accumulation of photosynthetic products from source organs (e.g., leaves and stalks) to reservoir organs (e.g., developing seeds). Therefore, delayed senescence helps to increase photosynthetic rate, extend photosynthetic duration, and promote the accumulation of nutrients in the seed. This significantly increases crop yield [4,5,6]. Studies have shown that chlorophyll content in leaves at the flowering stage in wheat is significantly and positively correlated with the number of grains per spike and TGW [7]. In sorghum, delayed leaf and stalk senescence during the filling stage increases yield, improves grain quality, and enhances resistance to stunting [8]. In other words, plants with green-holding characteristics can maintain photosynthetic function longer after flowering. This prolongs the grain-filling period and increases biomass accumulation and final grain yield.

There is a close correlation between delayed senescence and chlorophyll stabilization in crop leaves [5]. Leaf senescence is the final stage of leaf development and is most notably marked by chlorophyll degradation [9,10,11]. Chlorophyll is the core pigment of photosynthesis and drives energy conversion by capturing light energy [12]. Studies have shown that leaf chlorophyll degradation metabolism is regulated by a variety of enzymes and genes [13], with key factors including chlorophyll b reductase (NYC1 and its homologous gene NOL), 7-hydroxymethyl chlorophyll a reductase (HCAR), stay-green proteins (SGR/NYE, SGR2/NYE2, SGRL), pheophytinase (PPH), PaO, and red chlorophyll catabolite reductase (RCCR). Among these enzymes, PaO, a member of the Rieske family of non-heme oxygenases [14,15], assumes a central catalytic function—it specifically catalyzes the demineralization of the porphyrin ring of demagnesyl chlorophyllate a (Pheide a), which specifically catalyzes the oxidative rupture of the porphyrin ring of demagnesyl chlorophyll a (Pheide a) to generate the linear tetrapyrrole intermediate red chlorophylline (RCC); subsequently, RCC is converted to the primary fluorescent chlorophyll degradation product (pFCC) by the action of RCC reductase (RCCR). This series of “PaO pathways,” consisting of PaO-dominated enzymatic reaction steps, has now been elucidated [13]. Two major regulatory sites exist in the chlorophyll degradation pathway, namely chlorophyllase (CLH) and the PaO/RCCR complex [16,17]. In addition, it has been demonstrated that several chloroplast proteins play important roles in the biodegradation of chlorophyll centered on PaO [14]. Information on PaO gene function has been published for maize [18,19], Arabidopsis thaliana (A. thaliana) [20,21,22], broccoli [23], oilseed rape [24], rice [25,26], and wheat [27]. Investigations on a variety of crops have found that overexpression of PaO genes accelerates leaf chlorophyll degradation and senescence processes. These results provide a reference for the study of PaOs genes in sorghum, while the close association between senescence genes and yield provides a research basis for further excavation of superior haplotypes of sorghum PaOs genes associated with yield enhancement.

In sorghum genetic breeding, assessment of genetic diversity for leaf senescence and yield traits based on morphological and physiological traits is aimed at identifying genotypic variation in breeding value. The core objective is to screen optimal genotypes for crosses to produce progeny populations with sustained genetic gain [28,29]. Single-nucleotide polymorphisms (SNPs) are the most abundant genetic variants in plant genomes and are widely used in gene targeting, molecular marker breeding, and functional genomics studies [30,31,32,33,34]. Resolving nucleotide polymorphisms in conserved regions and constructing haplotypes can provide precise targets for analyzing the mechanism of inheritance of complex traits and molecular design breeding [35,36,37,38]. And KASP marker technology has become a representative method for SNP genotyping due to its rapid analysis and visual detection, providing a means to screen superior haplotype materials [39,40,41]. The development of KASP markers is currently being used in a large number of applications in crop genetic improvement breeding, such as wheat [42], maize [43], foxtail millet [44], rice [45], and potato [46]. Yu et al. [47] developed KASP markers Cold1-1-kasp and Cold1-2-kasp for the rice Cold1 gene to distinguish the three Cold1 gene haplotypes, Hap1, Hap2, and Hap3, which are associated with cold hardiness at the bud stage. The labeling results accurately distinguished the corresponding haplotypes. Shoot cold hardiness was improved by selecting rice germplasm for the Cold1 Hap2 genotype. In sorghum, Zhang et al. [48] developed KASP markers for nine alleles of the Tan1 and Tan2 genes associated with sorghum tannins, successfully differentiated between tannin and non-tannin cultivars, and characterized haplotypes, which provided a tool for tannin germplasm screening and molecular breeding of sorghum. Mining of superior allelic loci and haplotype analysis can optimize parental selection strategies, while molecular markers can assist in the screening of germplasm carrying target haplotypes.

The functional characterization of the PaO gene family has been studied in various plants, including A. thaliana, rice, maize, wheat, and tomato. However, as an important candidate gene resource, its molecular characterization, functional properties, and genetic variation are still unknown in sorghum. In this study, we performed the first genome-wide characterization of the sorghum PaO gene family (SbPaOs) and systematically resolved its chromosomal localization, gene structure, phylogeny, protein network interactions, GO annotation, and expression pattern using bioinformatics methods. qRT-PCR analysis further verified the correlation between SbPaOs and leaf senescence. In view of the close relationship between leaf senescence and yield, this study combined 250 sorghum germplasm resources and their resequencing data to analyze the natural variation of SbPaOs, revealing that their haplotypes were significantly associated with yield traits. By identifying superior haplotypes, we developed KASP markers for precision breeding and verified their accuracy in 96 sorghum germplasm. Ultimately, three superior haplotypes were identified that were significantly associated with TGW, and these haplotypes showed positive aggregation effects. The KASP markers developed for these three superior haplotypes provide an effective tool for molecular marker-assisted selection breeding in sorghum. This study lays the groundwork for an in-depth examination of the PaO gene family and provides valuable genetic resources for high-yield sorghum breeding.

2. Materials and Methods

2.1. Plant Material and Treatments

The 250 sorghum germplasm materials selected for this study were provided by the Fodder Sorghum Research Laboratory of the Sorghum Research Institute of Shanxi Agricultural University. The population integrates local varieties formed through long-term natural and artificial selection with regional characteristics, as well as a large number of modern breeding varieties. It covers parental lines that have made important contributions in the history of sorghum breeding in China, parents of large-scale popularized hybrids, and breeding materials selected independently by our team (Table S1).

Crops were planted for three consecutive seasons from 2022 to 2024 at the breeding base of the Sorghum Research Institute of Shanxi Agricultural University, Jinzhong City, Shanxi Province, China (37°41′12.0″ N, 112°44′18.0″ E): E1 (2022 Dongbai), E2 (2023 Dongbai), and E3 (2023 Dongbai). Before sowing, sorghum seeds of uniform size, color, and shape were selected and were fully mature, and seeds that were not suitable for planting were removed to ensure the germination rate of the seeds. The conditions were as follows: sowing row spacing: 55 cm, plant spacing: 20 cm, plot size: 3 rows per material and 5 m per row, replications: 3, and experimental design: randomized zone group test. The seeding of the identified materials was completed on the same day, and the planting density of the materials was the same; the depth of the seeding was 3~5 cm, the seeds were sown in strips, and the seeds were filled with pressure in time after sowing. Seedling: 3~4 leaves for inter-seedling, 5~6 leaves for seedling fixing. As far as possible, seedlings were fixed to maintain equal spacing. Only strong and positive seedlings were retained, and double-planted seedlings were removed. Seedlings were set within two days at each planting site, followed by consistent watering and fertilization, and after spiking, plants of uniform height were selected for bagging. Nine plant materials were selected from each variety in the experimental field for agronomic traits. To evaluate TGW, automatic seed analysis and a thousand-seed weighing instrument were used to measure 300~400 seeds, and the data of three groups were repeated and averaged; TGW was accurate to 0.01 g. The data of the three groups were repeated and averaged.

2.2. Identification of PaO Family Members in Sorghum

To accurately and comprehensively characterize the sorghum PaO gene family, we retrieved the sorghum genome-wide data from the Ensembl Plants database (http://plants.ensembl.org/Sorghum_bicolor/Info/Index/, accessed on 1 March 2024), including the annotation file (Sorghum_bicolor_NCBIv3.54.gff3) and the protein sequence file (Sorghum_bicolor_NCBIv3.pep.all.fa), to construct a local protein database. Hidden Markov Model (HMM) profiles of the PaO family (PF00355: Rieske [2Fe-2S] domain and PF08417: Pheophorbide a oxygenase) were employed to identify candidate sequences. Using the Simple HMM Search function in TBtools, we screened the local database for proteins containing both the Rieske [2Fe-2S] domain and the Pheophorbide a oxygenase domain, with an E-value threshold of 1 × 10⁻⁵ to remove redundant sequences. The resulting sequences were considered candidate members of the PaO gene family.

Conserved domains were verified using SMART (http://smart.embl-heidelberg.de/, accessed on 1 March 2024) and the NCBI Conserved Domain Database (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi/, accessed on 1 March 2024). The physicochemical properties of the identified SbPaO proteins were predicted with ExPASy-ProtParam (https://web.expasy.org/protparam/, accessed on 1 March 2024). Chromosomal locations were visualized using MG2C (http://mg2c.iask.in/mg2c_v2.0/, accessed on 1 March 2024), and gene structures were illustrated with the Gene Structure Display Server (GSDS, http://gsds.cbi.pku.edu.cn/, accessed on 1 March 2024). Conserved motifs were analyzed with Multiple Em for Motif Elicitation (MEME 4.11.4) suite (https://web.mit.edu/meme_v4.11.4/share/doc/release-notes.html, accessed on 1 March 2024), where the number of motifs was set to 10 and other parameters were kept at default values. Subcellular localization predictions were performed using WoLF PSORT (https://wolfpsort.hgc.jp/, accessed on 1 March 2024).

2.3. Sequence Alignment and Phylogenetic Analysis

Multiple sequence alignment of the full-length PaO protein sequences was performed using Clustal W with default parameters. The sequences included 5 SbPaOs, 3 AtPaOs, 7 OsPaOs, 5 ZmPaOs, 5 SiPaOs, and 6 CqPaOs. A phylogenetic tree was subsequently constructed with MEGA11 using the maximum likelihood method and 1000 bootstrap replicates. The final tree was visualized and graphically edited in FigTree v1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/, accessed on 2 October 2025).

2.4. Collinearity Analysis of the PaO Family

Chromosomal locations of the SbPaO genes were obtained from the Phytozome database. All non-redundant SbPaO genes were mapped onto the 10 sorghum chromosomes using TBtools v2.331. Gene duplication events among the SbPaO genes were analyzed using the Multiple Collinearity Scan toolkit (MCScanX). Syntenic relationships of PaO genes across sorghum, maize, millet, rice, and A. thaliana were visualized with the Multiple Synteny Plot function in TBtools v2.331.

2.5. Analysis of PaO Protein Interaction in Sorghum

To investigate the protein–protein interactions (PPIs) of SbPaOs in sorghum, the sequences of five PaO proteins were submitted to the STRING database (https://cn.string-db.org/, accessed on 1 March 2024) using sorghum as the reference species for constructing the PaO protein interaction network. The resulting network was visualized using Cytoscape software (version 3.10.0). Furthermore, proteins interacting with SbPaOs were subjected to Gene Ontology (GO) annotation, GO enrichment analysis, and KEGG pathway enrichment analysis through the BMKCloud platform.

2.6. Expression Analysis of SbPaOs

Based on our previously obtained transcriptome sequencing data, RNA-seq datasets were retrieved from the following tissues and developmental stages: roots, stems, leaves, glumes, anthers, embryos, primary branching differentiation stage (Ear 1), secondary branch differentiation stage (Ear 2), early stages of spikelet and floret differentiation (Ear 3), spikelet differentiation period (Ear 4), floret differentiation period (Ear 5), as well as seeds at 5, 9, 13, 17, 21, and 24 days after flowering (Seed 1, Seed 2, Seed 3, Seed 4, Seed 5, and Seed 6). The expression levels of SbPaOs genes, measured in fragments per kilobase of transcript per million (FPKM), were used to generate a heatmap, which was visualized using TBtools.

2.7. Total RNA Extraction and Real-Time PCR Analysis

Seeds of the sorghum cultivar L17R were cultivated in a greenhouse under controlled conditions: temperature of 26 ± 2 °C, a photoperiod of 14 h light/10 h darkness, and relative humidity of 65%. Root, stem, and leaf samples were collected at the three-leaf and one-heart stage. In parallel, plants were grown until grain formation at the Sorghum Research Institute breeding base of Shanxi Agricultural University for corresponding tissue sampling, aimed at analyzing the expression patterns of the SbPaOs family members in different tissues. For expression analysis under natural senescence conditions, leaf samples were taken at heading, flowering, 7 days after flowering (7 DAF), 14 DAF, 21 DAF, and 28 DAF. Using a 6 mm diameter punch, six leaf discs were collected from the upper, middle, and lower parts of the penultimate leaf at each time point.

Total RNA was extracted from various tissues and senescence-stage samples using the RNAprep Pure Polysaccharide and Polyphenol Plant Total RNA Kit (Tiangen Biotech (Beijing) Co., Ltd., Beijing, China) according to the manufacturer’s instructions. RNA purity was assessed based on the OD260/OD280 ratio (between 1.8 and 2.0), and integrity was verified by 1.2% agarose gel electrophoresis. cDNA was synthesized from RNA using the FastKing cDNA First-Strand Synthesis Kit (FastKing gDNA Dispelling RT SuperMix; Tiangen) following the manufacturer’s protocol. Gene-specific primers for SbPaO1 to SbPaO5 were designed in accordance with qRT-PCR primer design guidelines. Expression levels were quantified using SYBR^® Premix Ex Taq™ GC (Perfect Real Time; Takara Bio Inc., Japan) on a real-time PCR system. The 20 μL reaction mixture consisted of 10 μL TB Green Premix Ex Taq II, 0.4 μL of 10 μM forward/reverse primer mix, 1 μL cDNA template, and ddH₂O added to a final volume of 20 μL. The amplification program was as follows: 95 °C for 2 min; 45 cycles of 95 °C for 20 s, 60 °C for 20 s, and 72 °C for 20 s. The SbEIF4a gene (Sobic. 004G039400) was used as an internal reference to normalize the cDNA templates. All experiments included three biological replicates, and relative gene expression was calculated using the 2^−ΔΔCT method. The sequences of the primers used for qRT-PCR are listed in Supplementary Table S2.

2.8. Haplotype Analysis

Based on the resequencing data previously generated by our research group, we focused on the SbPaOs genes and extracted all SNP genotype information within the gene regions from the variant call format (VCF) files derived from the resequencing data. Specifically, the R package vcfR (version 1.15.0) was used to read and process the VCF files. Haplotype typing was performed according to the combinations of SNP genotypes within the target gene regions [49]. The association between SbPaOs haplotypes and thousand-kernel weight was analyzed using GraphPad Prism 8 with Duncan’s multiple range test, with a significance threshold of p < 0.05. In subsequent analyses, haplotypes represented by fewer than five accessions were defined as rare variants and excluded from further study.

2.9. DNA Extraction and KASP Marker Development

Genomic DNA was extracted from leaves of 96 sorghum seedlings using the CTAB method. The concentration and purity of the extracted DNA were measured with a microspectrophotometer. All samples had concentrations greater than 25 ng/μL and were diluted to a working concentration before being transferred into a 96-well PCR plate for high-throughput amplification. Flanking sequences of 150 bp on each side of the target SNP were used to design KASP assay-specific primers with the online tool Primer3 (version 0.4.0; primer sequences are provided in Supplementary Table S3). The PCR reaction mixture consisted of the following components: KASP 2× Master Mix (LGC Genomics, Product Code: KBS-0024-001, Teddington, Middlesex, TW11 0LY, UK), SNP Primer Mix, DNA template (25–100 ng), and nuclease-free water. The total reaction volume was 5.0 μL, comprising 2.5 μL of KASP 2× Mix, 0.07 μL of SNP Primer Mix, 1–2 μL of DNA template, and nuclease-free water, for a final volume of 5 μL. The thermal cycling conditions were as follows: initial denaturation at 95 °C for 10 min; 10 touchdown cycles of 95 °C for 20 s and 61–55 °C for 45 s (decreasing by 0.6 °C per cycle); followed by 34 additional cycles of 95 °C for 20 s and 55 °C for 45 s. After amplification, fluorescence was measured using a real-time PCR instrument (C1000 Touch™ Thermal Cycler, Bio-Rad, Hercules, USA). The KASP genotyping system uses FAM and HEX fluorophores to distinguish between the two alleles, with ROX serving as an internal reference for normalizing well-to-well variations in signal.

3. Results

3.1. Genome-Wide Identification of PaOs Genes in Sorghum

To identify PaO-related proteins in sorghum, we searched the sorghum genome database using Hidden Markov Models (HMMs) of the PaO family, including the Rieske [2Fe-2S] domain (PF00355) and the Pheophorbide a oxygenase domain (PF08417). A total of 11 full-length protein sequences were initially identified. After secondary screening using the SMART and CD-search databases, five members of the sorghum PaO gene family were ultimately confirmed and designated as SbPaO1 to SbPaO5. Among these, three genes were located on chromosome 1, while the other two were distributed on chromosomes 3 and 4, respectively (Figure 1B). Bioinformatic predictions revealed that the encoded proteins ranged from 524 to 542 amino acids in length, with molecular weights ranging from 58.54 kDa (SbPaO1) to 60.98 kDa (SbPaO5). The predicted isoelectric (pI) points ranged from 7.92 to 8.74. Since all pI values exceeded 7, these proteins were classified as typical alkaline proteins (Table S4). Subcellular localization predictions indicated that all five SbPaO family members are localized to the chloroplast (Table S4).

3.2. Motif Identification and Gene Structure Analysis of SbPaOs

Gene structure and protein conserved regions are important bases for studying gene function. To explore the gene structure of the SbPaO gene family members, we obtained the untranslated regions (UTRs) and CDS sequences of five SbPaO family members from the sorghum genome files and analyzed their exon numbers and UTRs. The results showed that SbPaO4 contained the highest number of exons (9), while SbPaO5 had the lowest (3). SbPaO1 and SbPaO2 contained 6 exons, and SbPaO3 contained 7. SbPaO2 contained 6 exons, while SbPaO3 contained 7. In addition, all four SbPaO genes except SbPaO2 contain UTRs (Figure 1A).

To investigate the structurally conserved sequences of the SbPaO protein family, we predicted five conserved motifs using MEME software (Table S5). Motifs 2 and 3 were present in all SbPaO proteins and were identified as functionally conserved domains, corresponding to the Rieske [2Fe-2S] domain (including the typical Rieske (2Fe-2S) and mononuclear iron-binding domains) and the Pheophorbide a oxygenase domain, respectively. Motifs 1 and 5 were also conserved and shared among all members. In contrast, Motif 4 was only detected in SbPaO1, SbPaO2, and SbPaO3 (Figure 1A).

3.3. Phylogenetic Analysis of the PaOs Gene Family in Sorghum

To investigate the phylogenetic relationships of SbPaOs with those from foxtail millet, maize, rice, quinoa, and Arabidopsis, we constructed a phylogenetic tree using the maximum likelihood method (bootstrap = 1000) (Figure 1C, Table S6). All PaO members from the different plants clustered into five subgroups (Group I–V). The five SbPaOs from sorghum were distributed across four groups: Group I, Group II, Group III, and Group V. Specifically, SbPaO1 was located in Group III, SbPaO2 and SbPaO3 were in Group V, SbPaO4 was in Group I, and SbPaO5 was in Group II. The phylogenetic structure indicated that SbPaOs are most closely related to monocotyledonous species such as maize and foxtail millet, while they exhibit a more distant relationship with the dicotyledonous model plant A. thaliana. Multiple sequence alignment revealed that the five SbPaO proteins share an average sequence identity of 40.73%, and all retain the highly conserved Rieske [2Fe-2S] domain (consensus pattern CxH…xxC). This domain was completely conserved across all members, confirming its essential role in the core catalytic function of chlorophyll ring opening (Figure S1).

3.4. Chromosomal Location and Gene Duplication of PaOs Genes in Sorghum

To investigate the patterns of gene duplication, loss, and rearrangement events in sorghum PaO genes and to clarify the origin and expansion mechanisms of this gene family, we performed a collinearity analysis of the sorghum PaO genes. The results indicated that no gene duplication events occurred within the sorghum genome for the SbPaO genes, suggesting functional conservation of this gene family. To further elucidate the evolutionary relationships of PaO genes across different plant species, a comparative collinearity analysis was conducted between sorghum and three monocot species (maize, foxtail millet, and rice) as well as one dicot species (A. thaliana) (Figure 1D, Table S7). The analysis revealed the presence of 4, 4, 3, and 1 homologous PaO gene pairs between sorghum and maize, foxtail millet, rice, and A. thaliana, respectively. These correspond to 4 sorghum genes paired with 4 maize genes, 4 sorghum genes with 4 foxtail millet genes, 3 sorghum genes with 3 rice genes, and 1 sorghum gene with 1 A. thaliana gene. Statistical analysis showed that the number of homologous gene pairs between sorghum and the monocot species (maize, foxtail millet, and rice) was significantly greater than that between sorghum and the dicot species (A. thaliana), further supporting a closer phylogenetic relationship of PaO genes among monocots. Moreover, SbPaO1, SbPaO3, and SbPaO5 were found to have collinear genes across all three graminaceous species, suggesting that these three genes may have been conserved during the evolution of graminaceous crops.

3.5. Sorghum PaO Protein Interaction Analysis

To further investigate the PPI of the sorghum SbPaOs gene family members, we constructed a regulatory network and a PPI network using the STRING protein database (Figure 2A). The results revealed that SbPaO1 interacts with Sobic.003G010600 (involved in porphyrin and chlorophyll metabolism) and Sobic.003G333800 (NADH–cytochrome b5 reductase 1), among others. SbPaO2 was found to interact with Sobic.003G333800 (NADH–cytochrome b5 reductase 1) and Sobic.008G038801 (endoplasmic reticulum chaperone BiP). SbPaO3 exhibited interactions with Sobic.003G333800 (NADH–cytochrome b5 reductase 1), Sobic.003G435800 (chlorophyllide a oxygenase, chloroplastic), Sobic.001G372600 (divinyl chlorophyllide a 8-vinyl-reductase, chloroplastic), Sobic.003G010600 (probable chlorophyll(ide) b reductase NYC1, chloroplastic isoform X3), and Sobic.006G268200 (protochlorophyllide reductase). SbPaO4 interacted with Sobic.003G333800 (NADH–cytochrome b5 reductase 1), Sobic.006G268200 (protochlorophyllide reductase), and Sobic.003G010600 (probable chlorophyll(ide) b reductase NYC1, chloroplastic isoform X3), among others. SbPaO5 was predicted to interact with Sobic.009G178200 (NADH–cytochrome b5 reductase 1) and Sobic.004G266400 (monodehydroascorbate reductase (NADH)).

The PPI network analysis indicated that SbPaO3 and SbPaO4 serve as hub genes with central roles in the network, exhibiting higher connectivity. Functional enrichment analysis of these interacting genes revealed significant GO biological processes related to the chlorophyll catabolic process, chlorophyll biosynthetic process, regulation of the tetrapyrrole metabolic process, and chlorophyll metabolic process. KEGG pathway enrichment highlighted the “Porphyrin and chlorophyll metabolism” pathway as the most significantly enriched (Figure 2B,C, Table S8). This study demonstrates that members of the sorghum PaO gene family—particularly the core hub genes SbPaO3 and SbPaO4—participate in regulating chlorophyll biosynthesis and degradation metabolism through specific protein interaction networks.

3.6. Tissue-Specific Expression Patterns of SbPaO Genes

To investigate the tissue-specific expression patterns of the sorghum PaO gene family, we analyzed previously obtained RNA-seq data from various tissues and developmental stages, including root, stem, leaf, glume, anther, embryo, Ear 1, Ear 2, Ear 3, Ear 4, Ear 5, and seeds at 5, 9, 13, 17, 21, and 24 days after flowering (Seed 1 to Seed 6). The results revealed that SbPaO2 was not expressed in any of the examined tissues, while the remaining SbPaO genes exhibited distinct spatiotemporal expression patterns (Figure 3A). Except for SbPaO2, all other SbPaO genes showed significantly higher expression levels in leaves compared to other tissues. SbPaO1, SbPaO3, SbPaO4, and SbPaO5 were specifically expressed in glume, stem, and Seed3. Additionally, SbPaO3 and SbPaO4 were expressed during various seed developmental stages (Seed 1 to Seed 6) and in the embryo.

To further determine the tissue-specific and developmental expression profiles of the SbPaOs genes, we performed qRT-PCR to analyze their expression in sorghum root, stem, leaf, and grain tissues, as well as in leaves collected at the heading stage, flowering stage, and 7, 14, 21, and 28 days after flowering (DAF). The qRT-PCR results demonstrated that SbPaO2 was not expressed in any of the examined tissues, which is consistent with the RNA-seq data (Figure 3B). In contrast, SbPaO1, SbPaO3, SbPaO4, and SbPaO5 exhibited the highest expression levels in leaves, followed by the stem, grain, and root. These findings further validated the accuracy of the RNA-seq results. Analysis of the relative expression levels of SbPaO1, SbPaO3, SbPaO4, and SbPaO5 during natural senescence indicated that their transcripts were most abundant in the late senescence stages, showing a consistent upward trend as senescence progressed and reaching peak levels at 28 DAF (Figure 3C). Based on the tissue-specific expression patterns of the PaO gene family in sorghum, we speculate that these genes may play crucial roles in leaf development and function. Furthermore, their upregulated expression during natural leaf senescence suggests that SbPaOs are importantly involved in the regulation of leaf senescence.

3.7. Haplotype Analysis of SbPaOs

Leaf senescence is closely associated with stay-green traits and grain yield, and previous studies have confirmed that SbPaO1, SbPaO3, SbPaO4, and SbPaO5 are highly expressed in leaves with expression levels positively correlated with senescence progression. To decipher genetic variations related to yield in these four genes, we conducted haplotype analysis based on resequencing data from 250 sorghum accessions. The frequency distribution of TGW over three growing seasons followed a normal distribution across environments (Figure S2). Association analysis between haplotypes and yield traits identified superior haplotypes for yield improvement. Specifically, for SbPaO1, five haplotypes were found, namely hap1 (143 accessions), hap2 (55), hap3 (20), hap4 (15), and hap5 (6), with TGW values of 27.35 g, 27.52 g, 28.68 g, 30.35 g and 23.36 g, respectively; hap4 showed the highest TGW and was considered the superior haplotype. For SbPaO3, five haplotypes were identified, including hap1 (14), hap2 (190), hap3 (15), hap4 (10), and hap5 (14), with TGW values of 23.75 g, 27.67 g, 26.85 g, 27.30 g, and 30.23 g, respectively; hap5 had the highest TGW and was the superior haplotype. For SbPaO4, four haplotypes were detected, specifically hap1 (9), hap2 (131), hap3 (88), and hap4 (9), with TGW values of 25.69 g, 27.54 g, 27.45 g, and 32.16 g, respectively; hap4 displayed the highest TGW and was the superior haplotype. For SbPaO5, five haplotypes were analyzed, comprising hap1 (70), hap2 (32), hap3 (78), hap4 (33), and hap5 (14), with TGW values of 26.93 g, 27.13 g, 27.84 g, 27.49 g, and 27.91 g, respectively; no significant differences were observed, indicating no superior haplotype. Overall, superior haplotypes SbPaO1-hap4, SbPaO3-hap5, and SbPaO4-hap4 were significantly associated with higher TGW (Figure 4), providing valuable genetic resources for molecular marker-assisted breeding aimed at improving sorghum yield.

3.8. Improvement of TGW Through Superior Haplotype Aggregation Effect

To analyze the potential of SbPaO1-hap4, SbPaO3-hap5, and SbPaO4-hap4 dominant haplotypes to increase TGW by polymerization, we further analyzed the effect of their polymerization in a population of 250 sorghum plants. SbPaO1-hap4, SbPaO3-hap5, and SbPaO4-hap4 increased TGW by 10.69%, 10.57%, and 17.20%, respectively (Figure 5). The polymerization results showed (Figure 6) that polymerizing the favorable alleles of SbPaO1-hap4 and SbPaO3-hap5 increased TGW by 21.89%, polymerizing the favorable alleles of SbPaO1-hap4 and SbPaO4-hap4 by 24.75%, and polymerizing the favorable alleles of SbPaO3-hap5 and SbPaO4-hap4 increased TGW by 18.78%, and polymerization of favorable alleles of three genes, SbPaO1-hap4, SbPaO3-hap5, and SbPaO4-hap4, resulted in a significant increase in TGW by 29.09%. These findings demonstrate that pyramiding the superior haplotypes of SbPaO1-hap4, SbPaO3-hap5, and SbPaO4-hap4 can effectively enhance the thousand-kernel weight in sorghum, thereby providing a novel strategy for molecular design breeding aimed at improving yield-related traits and highlighting the breeding potential of these superior haplotypes.

3.9. Development of KASP Markers for SbPaO1, SbPaO3, and SbPaO4

To rapidly screen germplasm materials carrying superior genotypes, KASP markers were developed targeting the superior haplotypes SbPaO1-hap4, SbPaO3-hap5, and SbPaO4-hap4. The SbPaO1 gene is located at 3,545,374 bp on chromosome 1, harboring a G-to-A mutation that distinguishes two genotypes: SbPaO1-G (hap1, hap2, hap3, and hap5) and SbPaO1-A (hap4). The SbPaO3 gene is located at 77,312,681 bp on chromosome 1, with a T-to-C mutation dividing it into SbPaO3-T (hap1, hap2, hap3, and hap4) and SbPaO3-C (hap5). The SbPaO4 gene is located at 73,704,614 bp on chromosome 3, containing a G-to-A mutation that separates it into SbPaO4-G (hap1, hap2, and hap3) and SbPaO4-A (hap4). Among these, SbPaO1-A (hap4), SbPaO3-C (hap5), and SbPaO4-A (hap4) represent the target genotypes. For the three SNP loci mentioned above, KASP markers SbPaO1-kasp, SbPaO3-kasp, and SbPaO4-kasp were developed and used for genotyping identification in 96 sorghum germplasm accessions. Through PCR amplification and fluorescence detection, the genotypes at SbPaO1_3,545,374, SbPaO3_77,312,681, and SbPaO4_73,704,614 could be determined based on fluorescence signals: red signals (HEX) represented homozygous AA, CC, and AA genotypes, respectively, while blue signals (FAM) represented homozygous GG, TT, and GG genotypes, respectively (Figure 7). Based on the genotyping results of these three KASP markers, materials carrying the target superior haplotypes were successfully identified.

To validate the accuracy of the markers, the KASP genotyping results were compared with the original haplotype data of the 96 materials (Table S9). The results showed that the SbPaO1-kasp marker detected 7 materials carrying the A allele and 89 carrying the G allele; the SbPaO3-kasp marker identified 10 materials carrying the C allele and 86 carrying the T allele; the SbPaO4-kasp marker found 9 materials carrying the A allele and 87 carrying the G allele. The genotyping results of all three KASP markers were completely consistent with the original SNP information at the corresponding loci, demonstrating that the SbPaO1-kasp, SbPaO3-kasp, and SbPaO4-kasp markers can effectively identify key genetic variations of the target genes with 100% accuracy.

4. Discussion

In light of global food security and nutritional demands, sorghum demonstrates significant potential as a core ingredient for novel healthy food products, owing to its exceptional drought tolerance, thermal adaptability, and rich functional components [50]. The extensive genetic diversity of sorghum provides critical germplasm resources for addressing food security and utilizing marginal lands; integrating genomic research with functional gene analysis, coupled with strategies for aggregating favorable alleles, will accelerate the designed breeding of ideal sorghum varieties with multiple target agronomic traits [51]. However, excessively rapid and premature leaf senescence shortens the duration of photosynthesis, directly impairing plant growth and crop yield. Therefore, an in-depth investigation of candidate genes associated with chlorophyll metabolism in sorghum is of great importance for enhancing grain yield. Numerous gene families have been identified in the sorghum genome, such as the MADS-box gene family [52], the GT47 gene family [53], BURP domain genes [54], U-box E3 ubiquitin ligase genes [55], the B3 Transcription Factor Family [56], and the C2H2 Zinc Finger Gene Family [57]. These studies have laid a foundation for a deeper understanding of sorghum classification, evolution, and functional gene extraction. The PaO gene family is a key gene family in plants involved in the chlorophyll degradation pathway, with its core function being the regulation of chlorophyll catabolism. It plays crucial roles in various biological processes, including plant growth, development, and environmental adaptation [13]. PaO genes have been successively identified in A. thaliana [20,21,22], maize [18,19], rice [25,26], wheat [27], and rapeseed [24], but have not yet been reported in sorghum. A total of five non-redundant PaO genes were identified in the sorghum genome, distributed across three chromosomes. However, the number of PaO genes in sorghum differs from that in other species, and no tandemly duplicated genes were found, indicating significant divergence in gene family expansion during evolution. All five SbPaO genes (SbPaO1–5) contain conserved Rieske [2Fe-2S] and Pheophorbide a oxygenase domains, suggesting that SbPaOs are highly conserved evolutionarily and may share similar functions. Subcellular localization predictions indicate that all five sorghum PaO genes are targeted to the chloroplast, consistent with previous studies, and function in leaf senescence and chlorophyll degradation [21]. Phylogenetic tree analysis revealed that sorghum PaO genes are most closely related to those of monocots, specifically maize and foxtail millet, while notably most distantly related to the dicot A. thaliana. Building on this, interspecific collinearity analysis showed four homologous pairs between sorghum and maize, four between sorghum and foxtail millet, three between sorghum and rice, but only one homologous pair was found between sorghum and A. thaliana. Together, these results indicate a high degree of conservation and functional similarity of PaO genes among graminaceous crops. Furthermore, the PPI network analysis of SbPaOs revealed that the interacting proteins are primarily enriched in biological processes and KEGG pathways related to chlorophyll metabolism. These findings suggest that the sorghum SbPaOs gene family likely plays a significant role in chlorophyll catabolism and is involved in chlorophyll degradation during leaf senescence in sorghum.

PaO genes exhibit tissue-specific expression patterns in different plants and play important roles in flower development, seeds, and leaves. For instance, BnPaO2 in rapeseed is expressed throughout seed development, while BnPaO1 is expressed during the early stages of seed development [24]. OsPaO in rice is highly expressed in leaves [20]. CaPaO in pepper is expressed in roots, stems, leaves, and flowers, with higher expression levels in leaves compared to roots, stems, and flowers [58]. In pear, DaPaO2, DaPaO3, and DaPaO8 genes are highly expressed in stems, while DaPaO5 and DaPaO6 are highly expressed in most tissues [59]. In this study, transcriptome data combined with qRT-PCR were used to analyze the expression of sorghum SbPaOs genes in different tissues. The results indicated that members of the SbPaOs family are expressed in roots, stems, leaves, and grains, with the lowest expression in roots and the highest in leaves, which is consistent with findings regarding OsPaO in rice. Notably, SbPaO2 showed no expression in any tissue, suggesting it may have undergone pseudogenization or is under strict regulatory suppression.

Furthermore, the function of PaO genes has been studied in many plants and is associated with senescence. For example, the expression of AtPaO in Arabidopsis is positively correlated with senescence, showing a 6.2-fold increase after 6 days of dark treatment [20]. OsPaO in rice and CaPaO in pepper are induced under natural senescence conditions [26,58]. The expression of BoPaO in broccoli exhibits a significant increase during postharvest senescence, correlating with chlorophyll degradation [23]. In this study, qRT-PCR analysis also demonstrated that SbPaO1, SbPaO3, SbPaO4, and SbPaO5 genes respond to natural senescence, with expression levels increasing over the senescence period. These results are consistent with the protein interaction analysis, indicating that SbPaO1, SbPaO3, SbPaO4, and SbPaO5 genes in the sorghum SbPaOs gene family play important roles in chlorophyll catabolism.

Leaf senescence regulates key physiological processes such as photosynthesis, nutrient transport, and stress response. Based on known regulatory mechanisms of senescence, researchers have developed various strategies to modulate its progression with the aim of improving crop yield [60,61]. In high-yield crop breeding, genetic loci associated with stay-green traits have been extensively accumulated in modern elite varieties [62,63,64,65]. Sorghum has developed rich genetic diversity through long-term evolution and domestication, making the mining of key genes and superior allelic variations for important traits crucial for breeding applications [66]. Single-nucleotide polymorphisms, as a core form of genomic variation, together with insertions-deletions (InDels) and copy number variations, regulate complex phenotypes [67,68,69]. Although SNP markers are widely used in breeding for the introgression of genetic variations, studies have shown that haplotype markers are significantly superior to single SNP markers in dissecting quantitative traits with low heritability and can effectively improve prediction accuracy [70,71,72,73,74]. This study confirmed that SbPaO1, SbPaO3, and SbPaO4 are significantly upregulated during leaf senescence, which is closely related to yield traits. Based on resequencing data from 250 sorghum accessions, we conducted haplotype analysis of the SbPaO1, SbPaO3, SbPaO4, and SbPaO5 genes. Three superior haplotypes significantly associated with TGW, namely SbPaO1-hap4, SbPaO3-hap5, and SbPaO4-hap4, were identified in SbPaO1, SbPaO3, and SbPaO4, increasing TGW by 10.69%, 10.57%, and 17.20%, respectively. Notably, germplasm carrying two or three of the superior haplotypes was identified in our sorghum panel. Analysis confirmed a positive additive effect, where the pyramiding of these haplotypes progressively enhanced TGW. The combination of SbPaO1-hap4 and SbPaO3-hap5 increased TGW by 21.89%, the combination of SbPaO1-hap4 and SbPaO4-hap4 increased TGW by 24.75%, the combination of SbPaO3-hap5 and SbPaO4-hap4 increased TGW by 18.78%, and the combination of SbPaO1-hap4, SbPaO3-hap5, and SbPaO4-hap4 significantly increased TGW by 29.09%. These results reveal the pyramiding effects of superior haplotypes in sorghum SbPAOs, providing support for their potential as high-yield molecular modules and thereby laying a theoretical foundation for multi-gene pyramiding breeding and pinpointing design targets. The primary value of functional markers lies in assisting breeders to precisely screen superior allelic variations in germplasm resources [75,76], which can significantly enhance the predictability of parental selection and the efficiency of progeny selection [77,78]. For the superior haplotypes SbPaO1-hap4, SbPaO3-hap5, and SbPaO4-hap4, specific KASP markers SbPaO1-kasp, SbPaO3-kasp, and SbPaO4-kasp were developed and validated in 96 accessions, confirming their ability to unambiguously distinguish different haplotypes and providing technical support for efficient screening of target genotypes. The superior haplotypes identified in this study can be used for the precise selection of optimal haplotype combinations controlling TGW, and the KASP markers serve as effective tools for genetic improvement and cultivar development in sorghum breeding populations. This study lays the groundwork for future breeding strategies by identifying potential routes for crop improvement. Future efforts may leverage these TGW-linked haplotypes for marker-assisted parent selection, to explore allele pyramiding through targeted crossing and selection, thereby contributing to the goal of breeding higher-yielding varieties.

5. Conclusions

Pheophorbide a oxygenase is a key enzyme in the chlorophyll degradation pathway and plays an important role in plant senescence, influencing photosynthetic efficiency and nutrient remobilization. This study presents the first systematic analysis of the PaO gene family in sorghum, identifying five SbPaO members. Examination of gene structure, conserved motifs, phylogenetic relationships, chromosomal distribution, and interspecies collinearity revealed their genomic characteristics and evolutionary patterns. The PPI network and GO annotation analyses confirmed the involvement of SbPaOs in chlorophyll catabolism. Expression profiling showed that SbPaO1, SbPaO3, SbPaO4, and SbPaO5 are most highly expressed in leaves and are responsive to natural senescence. Haplotype analysis identified three superior haplotypes—SbPaO1-hap4, SbPaO3-hap5, and SbPaO4-hap4—which were associated with increased TGW and exhibited significant positive pyramiding effects. Based on these haplotypes, KASP markers were developed, enabling rapid germplasm identification and targeted genetic improvement. This study fills a critical knowledge gap in the genomic and functional characterization of the PaO family and provides a foundation for further investigation of the Pheophorbide a oxygenase domain in sorghum and other plants. Our findings demonstrate the potential of pyramiding superior SbPaO haplotypes as a promising strategy for enhancing TGW in sorghum. Future work can build upon the findings of this study by using marker-assisted selection as guidance, employing these superior haplotypes to screen parental lines, and pyramiding favorable alleles through hybridization and directional selection, thereby ultimately developing high-yielding new cultivars.