chl phenotypic characterization
We developed an ethyl methane sulfonate mutant library using the wheat cultivar Kenong 9204 (wild type, WT) and identified a mutant with significant leaf color changes, accompanied by changes in chlorophyll (Fig. 1E–F) - chlorophyll (chl). The mutant exhibited a yellow-green leaf phenotype at seedling stages (Fig. 1A). In comparison to the WT, chl showed significant reductions in chlorophyll a (− 18%), chlorophyll b (− 72%), and Soil and Plant Analyzer Development (SPAD) levels (− 28%), while carotenoid levels were significantly increased (+ 53%) (Fig. 1E–H). Chlorophyll content changes have the potential to affect photosynthetic efficiency, while chlorophyll fluorescence parameters can be used to evaluate photosynthetic efficiency in plants from multiple perspectives (Wang et al., 2022). Specifically, chl exhibited significant reductions in minimum initial fluorescence (Fo) (− 26%), maximum fluorescence (Fm) (− 17%), steady-state fluorescence decay rates (rfd_lss) (− 17%), and non-photochemical quenching under steady-state light (NPQ_lss) (− 55%) when compared to WT. In contrast, a slight difference in the maximum photosynthetic efficiency of photosystem II (Fv/Fm) (+ 3.19%) was observed (Fig. 1B–D, I–M). These results indicated that chl-mediated reductions in photosynthetic pigments decreased chlorophyll fluorescence and plant vitality while having a relatively smaller impact on maximum energy conversion efficiency in the Photosystem II (PSII) reaction center.
To investigate the potential impact of the mutation on chl chloroplasts, we used transmission electron microscopy (TEM), which showed that chl chloroplasts exhibited significant deformations (Fig. 1N, O). Specifically, a notable reduction in the number of plastoglobulus during the seedling (− 71%) and jointing stages (− 25%) was observed when compared to WT (Fig. 1P). Similarly, a substantial reduction in stroma numbers was observed at both the seedling (− 58%) and jointing stages (− 58%) (Fig. 1Q).
chl displayed a yellow-green leaf phenotype throughout the entire growth period (Fig. 2A, B). To unravel the implications of this on yield-related traits, a more detailed investigation was conducted. Under field conditions, chl exhibited a significant reduction in various agronomic traits when compared to WT (Table S2). Specifically, tiller number (− 61%), plant height (− 36%), spike length (− 18%), spikelet number per spike (− 22%), awn length (− 11%), flag leaf length (− 21%), and flag leaf width (− 12%) were all reduced (Fig. 2C, F–I, Table S2). Notably, a reduction in the second internode length (− 45%) was more pronounced than the third (− 25%) (Table S2). Eventually, the decline in photosynthesis levels significantly declined the grain number per spike (-18%), grain length (− 12%), grain width (− 27%), thousand-grain weight (− 36%), and yield per plant (− 61%) (Fig. 1D, E, K–O).
Molecular cloning of the candidate gene TaCHLI-7D
To analyze the genetic factors responsible for chl phenotypic changes, we backcrossed chl with the WT and the wheat variety Yanong 999. Genetic analyses indicated that chl was a mutant controlled by a nuclear recessive gene (Table S3). For the two F3 segregating populations with extreme phenotypes (yellow or green), 30 plants were selected for bulked segregant exome sequencing (BSE-Seq). We observed a dense distribution of high-quality single nucleotide polymorphisms (SNPs) in the 575–595 Mb region of chromosome 7D (based on the Chinese Spring RefSeq v2.1 sequence) (Fig. 3A). From these findings, we used map-based cloning and designed 10 polymorphic insertion–deletion (InDel) markers based on InDel sites between 580–590 Mb. Genotyping of 392 yellow plants in the segregating population narrowed the target interval to a 5.09 Mb region in 582.7–587.8 Mb (Fig. 3B). We identified 21 genes and five SNPs in this interval (Table S4), with only SNP 1 located in the TraesCS7D03G110950 promoter region, which encoded a magnesium ion chelation subunit with roles in chlorophyll a and b biosynthesis pathways. Therefore, we hypothesized that TraesCS7D03G1109500 was a chl candidate gene. Domain analysis showed that this gene belonged to the CHLI superfamily, so it was named TaCHLI-7D (Fig. 3C). Sequence analysis of WT and chl revealed that TaCHLI-7D contained a SNP on the 58301094 base of chromosome 7D, leading to a protein alteration from Asp to Asn at position 187 (Fig. 3C). From these results, TaCHLI-7D was posited as a chl candidate gene and underwent further study.
TaCHLI-7D functional validation in wheat mutants and rice overexpression lines
To investigate TaCHLI function in wheat, we screened an EMS mutant library of the WT. We identified three mutants in TaCHLI: chl-7b-1 (Pro82Ser), chl-7b-2 (Ala297Thr), and chl-7d-1 (Gly357Glu) (Fig. S1A, B), which exhibited yellowing phenotypes similar to chl during seedling stages. These mutants displayed varying degrees of leaf chlorosis. In later growth stages, chl-7b-1 and chl-7d-1 demonstrated whole-plant or leaf senescence, while chl-7b-2 exhibited a yellowing phenotype similar to chl (Fig. 4C).
Compared with WT, chlorophyll a (− 46%, − 61%, and − 29%), chlorophyll b (− 42%, − 13%, and − 23%), and SPAD (− 10%, − 16%, and − 28%) levels in chl-7b-1, chl-7b-2, and chl-7d-1 mutants, respectively, all decreased, while carotenoid levels decreased and were unchanged in chl-7b-1 (− 45%) and chl-7b-2, respectively, but increased in chl-7d-1 (28%). Additionally, the agronomic traits of chl-7b-1, chl-7b-2, and chl-7d-1, such as plant height (− 33%, − 10%, and − 30%), spike length (− 26%, − 14%, and − 37%), flag leaf length (− 18%, − 26%, and + 1%), and flag leaf width (− 60%, − 42%, and − 22%), respectively, were all changed (Fig. 4C, J–M). Also, most yield traits such as grain length (− 24%, − 5%, and − 9%), grain width (− 33%, − 4%, and − 20%), thousand-grain weight (− 78%, − 33%, and − 36%), and yield per plant (− 95%, − 79%, and − 67%), respectively, were significantly reduced (Fig. 4D, E, N–Q). Chl-7b-1 experienced the most severe decline due to whole-plant senescence, resulting in smaller grains. Therefore, TaCHLI-7B/D mutations impacted on photosynthesis, leading to chlorotic or prematurely senescent phenotypes similar to chl and ultimately affecting yield traits.
To further explore TaCHLI-7D biological functions, we cloned full length TaCHLI-7D into an overexpression transgenic vector ZH11-OE (pUbi::TaCHLI-7D) and introduced it into the rice variety Zhonghua 11 (ZH11). Two different transgenic lines (ZH11-OE#1 and ZH11-OE#2) were selected for observations. Compared to ZH11, transgenic ZH11-OE plants exhibited significant increases in transcript levels, ranging from 7.78 to 14.53 times higher (Fig. S1E). In contrast to ZH11, the overall growth vigor of ZH11-OE plants was poorer (Fig. S1A), with shorter flag leaves (− 29% and − 5%, respectively) and lower SPAD (− 7% and − 18%, respectively) in both lines (Fig. S1B, E, F). Plant height (− 5% and − 2%, respectively), grain length (− 4% and − 5%, respectively), and grain width (− 4% and − 4%, respectively) did not change significantly (Fig. S1G–I). Eventually, this caused severe yield reductions, with thousand-grain weight (− 42% and − 32%, respectively) and yield per plant (− 50% and − 46%, respectively) decreased (Fig. S1C, K, L). Furthermore, germination rates in both the lines decreased by 47% (Fig. S1D, M). Therefore, TaCHLI-7D overexpression severely affected germination and photosynthesis, leading to reduced yields. From these observations, both wheat mutants and TaCHLI overexpression in rice indicated that the mutation affected chlorophyll biosynthesis, which severely affected SPAD values and yields, and was potentially valuable in improving photosynthetic traits in plants.
chl transcriptome sequence analysis
In an in-depth analysis of yellow phenotype formation in chl, we sampled first leaves and flag leaves at seedling and jointing stages to extract RNA for RNA-Seq, with clean reads aligned with the Chinese Spring RefSeq v2.1 genome. Differentially expressed genes (DEGs) were identified based on |log2 (fold change) | ≥ 1 and P < 0.05 values. Seedling stages generated 6135 DEGs, with 4032 down- and 2103 up-regulated genes (Fig. 5A). In contrast, jointing stages generated 8419 DEGs, with 4108 down- and 4311 up-regulated genes (Fig. 5B). Importantly, 1134 DEGs were shared between stages (Fig. 5C). We selected six common genes from both stages for quantitative real-time PCR (qRT-PCR), with results consistent with RNA-Seq trends (Fig. 5E).
DEG enrichment analyses were conducted across seedling and jointing stages, and DEGs common to both were analyzed using Gene Ontology (GO) terms. GO annotations segregated DEGs into three categories: biological processes, molecular functions, and cellular components, with notable enrichment in the former two (Table S5). At seedling stages, GO terms were predominant for the reductive pentose-phosphate cycle, jasmonic acid signaling regulation, and ribulose-bisphosphate carboxylase activity, along with pathways related to photorespiration (Fig. 5D). It was noteworthy that pathways related to cellular responses to heat and water stress were enriched, suggesting that TaCHLI-7D had potential in drought stress in wheat.
At jointing stages, chloroplast and photosynthesis-related pathways, including chloroplast organization and PSII, were predominantly enriched (Fig. 5D). This indicated that photosynthesis-related genes had crucial roles at this stage. Moreover, a combined DEG analysis from both stages revealed significant enrichment in photosynthetic electron transport in PSII, the chloroplast thylakoid lumen, and the magnesium-dependent protein serine phosphatase activity pathway, which were all related to photosynthesis and suggested that TaCHLI-7D was closely related to photosynthesis.
Phylogenetic and expression analysis of TaCHLI
To explore TaCHLI-7D functions in wheat, we performed a BLASTP analysis using its protein sequence against the international wheat genome sequencing consortium (IWGSC) RefSeq v2.1 database (Zhu et al., 2021). TaCHLI-7A (TraesCS7A03G1163900) and TaCHLI-7B (TraesCS7B03G1030900), which were highly homologous to TaCHLI-7D, were identified. TaCHLI-7A, TaCHLI-7B, and TaCHLI-7D protein lengths were 421 amino acids (aa), 420 aa, and 420 aa, respectively (Table S6), and all belonged to the CHLI superfamily, suggesting similar functions in wheat. Phylogenetic analysis revealed that TaCHLI homologs were divided into two groups in cereals, and that they were most closely related to barley (HORVU. MOREX. r3. 7HG0738240) and rice (Os03g0563300) (Fig. 6A).
Based on TaCHLI-7A/7B/7D sequences, specific qRT-PCR primers were designed to analyze their expression levels across different tissues. We observed that TaCHLI-7A/7B/7D were expressed in all sampled tissues, exhibiting higher expression in major photosynthesis tissues, such as spikes and leaves, and lower expression in roots and grains (Fig. 6B–D). Moreover, TaCHLI-7A/B/D expression in chl was higher in roots and stems, and lower in leaves, ears, and grains when compared with WT (Fig. 6B–D). Subcellular localization studies demonstrated that TaCHLI-7D was located in chloroplasts, which suggested it impacted on chloroplast development, consistent with TEM observations in chloroplasts (Fig. 6E, Fig. 1N, O).
TaCHLI haplotype analysis
To investigate natural variations in TaCHLI toward agronomic traits, coding and 2 kb promoter regions in TaCHLI-7A/B/D in 42 wheat accessions (with high genetic diversity) were analyzed for sequence polymorphisms (Table S7). For TaCHLI-7A, six co-segregated SNPs were identified and divided into two haplotypes. A kompetitive allele-specific PCR (KASP) marker (7A-KASP) was designed and used to detect TaCHLI-7A haplotypes, based on SNP 1 (Fig. 7A). Additionally, four TaCHLI-7B haplotypes were distinguished by two InDels and one SNP. Based on the previously identified InDel 1 and InDel 2, two InDel markers (TaCHLI-7B-1 and TaCHLI-7B-4) were developed and detected using polyacrylamide gels (Fig. 7I). Primers are listed (Table S1). However, no nucleotide variations were identified in TaCHLI-7D (Fig. 7O).
Next, we calculated the best linear unbiased estimate (BLUE) values for six agronomic traits across 314 wheat accessions in eight environments between 2020 and 2023. When combined with TaCHLI haplotypes and phenotypic data from multiple environments and BLUE values from 314 wheat accessions, TaCHLI haplotype effects were clarified. When compared with TaCHLI-7A-HapI, TaCHLI-7A-HapII reduced flag leaf length (− 7%), flag leaf width (− 5%), and spike stem length (− 16.57%), and increased thousand-kernel weight (+ 3%) and yield per plant (+ 5%), but plant height showed no significant differences (Fig. 6C–H). For TaCHLI-7B, when compared to the other three haplotypes, TaCHLI-7B-HapII showed a significant decrease in plant height (− 8%, − 10%, and − 11%) and spike stem length (− 17%, − 24%, and − 30%), respectively, while increased thousand-kernel weight (+ 3%, + 3%, and + 4%), and yield per plant (+ 4%, + 3%, and + 4%), respectively, were identified (Fig. 7K–N). These findings suggested TaCHLI-7A/B haplotypes could provide effective genetic resources and molecular markers for high-yield cultivar development.
To determine TaCHLI haplotype selection characteristics in wheat breeding, we assessed 200 Chinese wheat accessions from different breeding years (Xu et al., 2022) and used them to analyze frequency variations in different TaCHLI-7A/B haplotypes. TaCHLI-7B-HapII frequency was raised from landraces (%) to modern cultivars (%) (Fig. 7J), while the frequency of the two TaCHLI-7A haplotypes remained essentially unchanged (Fig. 7B). Moreover, we investigated the geographical distribution characteristics of TaCHLI haplotypes in 314 wheat accessions from six continents (Xu et al., 2022). For TaCHLI-7A, the TaCHLI-7A-HapII distribution frequency accounted for a larger ratio than TaCHLI-7A-HapI in six continents, which suggested that TaCHLI-7A-HapII varieties could widely adapt to complex conditions (Fig. S2A). For TaCHLI-7B, TaCHLI-7B-HapI/III/IV were widely distributed across all continents, while TaCHLI-7B-HapII was only found in Asia (20%) and North America (14%) (Fig. S2B). Thus, TaCHLI-7B-HapII was positively selected in wheat breeding processes and had great potential for global wheat breeding selection.