Background

The leaves are the main organs for photosynthesis and are vital for plant growth and development. Leaf traits, especially leaf color, are considered important morphological traits that are relevant to photosynthesis potential and grain yield. Leaf color is mainly determined by the content of chlorophyll and carotenoids, which are photosynthetic pigments. In higher plants, photosynthetic pigments fix light energy and convert it into chemical energy to synthesize carbohydrates. Chlorophyll is the primary photosynthetic pigment that harvests light and drives electron transport in reaction centers [1]. Leaf color mutants with photosynthetic pigment deficiency or abnormal chloroplast structures can serve as a subject for systematic investigation of the mechanisms of photosynthesis and photosynthetic pigment biosynthesis or degradation in plants [2].

Recently, leaf-color mutants of different crops or plants, such as rice [3,4,5,8], Arabidopsis thaliana [9], and Chinese cabbage [10], have been used to identify steps associated with chlorophyll metabolism or chloroplast development. Studies have elucidated the molecular mechanisms of leaf-color mutants, especially the yellowing phenotype, which is most commonly observed in higher plants. Yellow-leaf mutants usually have shorter plant height, lower grain yield, deficiency in photosynthetic pigments, and abnormal chloroplast ultramicrostructures compared with the wild type [7, 11,12,13,14]. For example, the yellowish-leaf mutant ydl of rice has photosynthetic pigment deficiencies, degraded chloroplasts, and disordered matrix [4]. Genes that regulate leaf color formation and chloroplast development have been identified in numerous plants [2, 4, 5, 8,9,10,11, 13]. More than 70 genes corresponding to leaf color variations have been cloned [15]. Mutations in the BC12/GDD1 gene resulted in low chlorophyll content and short plant height in rice, and a mutation in the EMBRYO DEFECTIVE (EMB1923) gene promoter is associated with chlorophyll deficiency in Chinese cabbage [4, 10]. A novel gene SiYGL1 in foxtail millet is responsible for a deficiency in chlorophyll content [8]. An incomplete dominant gene Y1718 in a yellow-green leaf mutant (Ygm) contributed to yellow, yellow-green, and green leaf phenotypes from the jointing to adult stage [7]. Virus-induced gene silencing of the Mg chelatase CHL1 and CHLD (CHL1 and CHLD) genes both resulted in a yellow leaf phenotype in pea [16]. A single base mutation in the AtCHL1 gene led to a yellow-green or white phenotype in Arabidopsis [17].

Moreover, both internal and external factors, such as temperature, salt, and biotic and abiotic stresses, can influence photosynthetic pigment content and chloroplast development through alterations in gene expression or post-translational modification of proteins involved in chlorophyll metabolism [3, 4, 18,19,20,21]. For example, in common vegetable crops, photosynthetic pigment metabolism and chloroplast development are disrupted under drought stress [22]. An Arabidopsis mutant deficient in stay-green2 (SGR2) exhibited early leaf yellowing under age-, darkness-, and stress-induced senescence conditions [9]. Chloroplast development is also controlled by several hormones, including brassinosteroids, cytokinins, auxins, and gibberellins, which control chloroplast development, particularly during the early stages of plant development [23]. Several BR-related mutants in Arabidopsis have fewer grana stacks than the wild type [24,25,26]. In addition, CTK promotes several steps in chlorophyll biosynthesis, including the formation of 5-aminolevulinic acid (ALA; the first step in the biosynthesis pathway of tetrapyrroles) and light-dependent conversion of protochlorophyllide (Pchlide) into chlorophyllide [27,28,29]. In Arabidopsis, 29 chloroplast-related proteins, including CAB (LHCB1), LSU, and LHCB2, respond to auxin treatment [8,9]. Leaf color mutants are ideal models for understanding photosynthetic pigment metabolism and functional chloroplast development.

Tolerance to poor-quality soil and drought as well as high photosynthetic rate are important features of broomcorn millet. To date, no leaf-color mutants or gene map** data have been reported for broomcorn millet. Therefore, we developed a yellow-leaf mutant (YX-yl) with ethyl methanesulfonate treatment, and a yellow-leaf phenotype was observed during the entire growth stage. In this study, physiological and biochemical analyses integrated with transcriptome profiling were performed to elucidate the molecular mechanisms of the leaf-color mutant and explore the candidate genes controlling the mutant phenotype. The results will lay material and molecular foundations for further analysis of the photosynthetic mechanism of C4 plants.

The yellow-leaf phenotype is closely associated with chlorophyll and carotenoid pigment biosynthesis

Leaf color variation is a common phenomenon in plants, and the main reason for an abnormal leaf color is an increase or decrease in chlorophyll content in the leaves. Our results revealed that the leaves of the YX-yl mutant exhibited much lower chlorophyll content, especially Chl b content, compared with the wild type, which played an important role in harvesting light and energy conversion. We hypothesized that the yellow phenotype of YX-yl leaves is caused by a deficiency in chlorophyll content, especially Chl b content, which has also been reported in cabbage [2], wheat [7], rice [13], foxtail millet [8], and Arabidopsis [9].

Chlorophyll and carotenoid biosynthesis are determined by complex biological processes, and blockage of any step in this process can lead to a decrease in chlorophyll or carotenoid content, which in turn results in a change in leaf color. For example, the synthesis of ALA, Proto IX, and Chl a/b is regulated by glutamine-tRNA reductase, Mg-chelatase, and chlorophyllide A oxidase, respectively. If the synthesis process is inhibited or the activities of these enzymes change, we can analyze the reason for the blockage of the biosynthetic process of chlorophyll according to the accumulation of chlorophyll precursors. A previous study implied that the absence of or damage to enzymes responsible for converting Chl a into Chl b in barley might be the reason for the decrease in Chl b content [37]. Many of the reported chlorotic mutants exhibit reduced chlorophyll biosynthesis because of the lower activity of Mg-chelatase [38,39,40]. In the present study, comparative analysis of precursor content in the leaves of YX and YX-yl plants revealed that the PBG content was much higher, whereas Proto IX, Mg-Proto IX, and Pchlide were lower in YX-yl. We inferred that the blockage of the step between PBG and Proto IX led to a decrease in chlorophyll content, finally resulting in a yellow-leaf phenotype.

The yellow-leaf phenotype is closely related to photosynthesis and chloroplast development

The photosynthetic rate and chloroplast ultrastructure are directly affected by leaf color mutations [4, 8, 13, 14, 41]. In a leaf-color mutant of rice, the chloroplasts are degrade, the matrix is disordered, and the thylakoids almost completely disappear [4]. In a novel leaf-color mutant of foxtail millet, disorganized thylakoids and considerably fewer starch granules have been observed [8]. Moreover, a green-yellow leaf mutant of wucai exhibited fewer chloroplasts per cell and looser stromal lamellae [41]. Similar results were found in the present study: compared with that in the wild type (YX), chloroplast development was inhibited in the YX-yl mutant owing to degeneration of the chloroplast ultrastructure, with fewer grana thylakoids and almost no stroma thylakoids, which might result in poor chloroplast function and accelerated leaf yellowing.

The lower content of photosynthetic pigments, combined with poorly developed chloroplasts, might be the reason for the poorer photosynthetic performance and light transfer efficiency through the photosynthetic electron transport chain. The results of both photosynthetic performance and fluorescence kinetic parameter analyses revealed that yellow color had a negative impact on the photosynthetic machinery of YX-yl leaves compared with the wild-type green leaves. In previous studies, a lower chlorophyll content was found to occur independently of the photosynthetic rate and light-use efficiency. For example, in foxtail millet, the leaf-color mutant maintains a relatively higher Pn and light-use efficiency [8]. Furthermore, in rice, the leaf-color mutant ygl7 has high light energy conversion efficiency and solar energy capture efficiency [8]. Our study revealed that the upregulated expression of PmGSAM and PmGluTR might be the reason for the higher PBG content. Although upregulated expression was observed for PmPBGD, PmUROD, and PmCPO, which encode key enzymes for chlorophyll biosynthesis, the content of Proto IX, Mg-Proto IX, and Pchlide severely decreased in the YX-yl leaf-color mutant, indicating that inhibition of the biosynthesis step between hydroxymethylbilane and uroporphyrinogen might be the major reason for the reduction in Proto IX and Mg-Proto IX content. Further studies should be conducted focusing on the genes or enzymes involved in this biosynthesis step. Only one downregulated gene was observed, namely PmCAO, which is involved in the biosynthesis step between Chl a and Chl b; therefore, we hypothesized that the decrease in Chl b content might be due to the downregulated expression of PmCAO. The PmVDE gene encodes a key enzyme in the lutein cycle in higher plants, which is considered the main mechanism to protect the photosynthetic apparatus from excess light energy. The development of chloroplast was inhibited in the YX-yl mutant, and from transcriptome analysis, we found that two genes (PmLHCP and PmPNPT) involved in chloroplast development were upregulated in the YX-yl mutant. These results showed that chlorophyll metabolism, chloroplast development, and chloroplast transcription were impaired in YX-yl plants, leading to chloroplast thylakoid degradation, reduced chlorophyll content, and yellowish leaves. Further research will be required to confirm which candidate genes play a vital role in chlorophyll metabolism or chloroplast development in YX-yl mutants and to explore the gene-dependent processes involved in these metabolic pathways.

Conclusions

Phenotypic, photosynthetic, and transcriptome analyses were conducted for yellow-leaf mutant (YX-yl) and wild-type (YX) broomcorn millet. The yellow-leaf phenotype had lower SWPP and GWPP values, which had negative effects on grain yield. The lower Chl a and Chl b contents, abnormal chloroplast structures, and lower photosynthetic performance in the leaves of YX-yl indicated that chlorophyll biosynthesis (biosynthesis step between PGB and Proto IX) and chloroplast development were partially inhibited. Furthermore, we performed transcriptome analysis using RNA-seq to elucidate the molecular mechanisms underlying the yellow-leaf phenotype. We identified nine DEGs related to chlorophyll metabolism and chloroplast development. Among these genes, six genes related to chlorophyll biosynthesis, including PmUROD, PmCPO, PmPBGD, PmGSAM, and PmGluTR, were upregulated, whereas PmCAO was downregulated. In addition, PmVDE and PmLHCP, which are related to chlorophyll metabolism and chloroplast development, were upregulated in YX-yl. The findings of this study provide molecular evidence for the development of leaf color and chloroplasts in broomcorn millet.

Methods

Plant materials

The broomcorn millet yellow leaf mutant YX-yl was isolated from the Yi Xuan Da Hong Mei cultivar (YX), which was extensively cultivated in China. YX seeds were mutagenized with 1.0% (v/v) ethyl methanesulfonate solution. One M2 line, carrying the yellow-green leaf phenotype, was identified. YX (wild type) and YX-yl (mutant type) were used as materials for phenotypic and genetic analyses in the present study.

Growth conditions and agronomic traits evaluated

For the YX cultivar (wild type) and YX-yl (mutant type), all plants were grown at the experimental station of Shanxi Agricultural University (Shanxi, China). Three uniformly developed plants were collected at the flowering stage or after harvest for phenoty**: plant height (PH), peduncle length (PL), panicle length (PaL), basal stem diameter (BSD), flag leaf length (FLL) and width (FLW), spike weight per plant (SWPP), and grain weight per plant (GWPP).

PH, PL, BSD, PaL, FLL, and FLW were measured at the flowering stage. PH was determined as the distance from the soil surface to the top of the spike. PL was measured from the base of the spike to the ligule of the flag leaf. BSD was determined as the midpoint of the basal stem from the soil surface. FLL and FLW were determined from the base to the top or midpoint of the flag leaf, respectively.

Yield components were investigated after harvest. Panicle length (PaL) was measured from the base to the top of the spike. SWPP was determined as the total spike weight per plant. The GWPP was determined as the total grain weight per plant.

Photosynthetic pigments and chlorophyll fluorescence

The leaves of YX and YX-yl were collected to measure the photosynthetic pigment content at the seedling, booting, heading, and mature stages, with three biological replicates tested for each sample. The leaves were obtained from the flag leaves without the main leaf vein. Approximately 0.2 g of fresh leaves was immersed in 95% ethanol for 48 h until the leaves turned white. The absorbance of each leaf extract was measured using a UV-1800 ultraviolet/visible spectrophotometer (UV-1800pc; MAPADA, Shanghai, China) at 470, 649, and 665 nm. Chlorophyll a (Chl a), chlorophyll b (Chl b), carotenoid, and total chlorophyll content was calculated according to the following formula:

$$\mathrm{Chl a }\left(\mathrm{mg}/\mathrm{g}\right)=\left({13.95\mathrm{A}}_{665}-{6.88\mathrm{A}}_{649}\right)\times 0.01/0.2$$
$$\mathrm{Chl b }\left(\mathrm{mg}/\mathrm{g}\right)=\left({24.96\mathrm{A}}_{649}-{7.32\mathrm{A}}_{665}\right)\times 0.01/0.2$$
$$\mathrm{Car }\left(\mathrm{mg}/\mathrm{g}\right)=\left({1000\mathrm{A}}_{470}-3.27\times \mathrm{Chl a }-104\times \mathrm{Chl b}\right)\times 0.01/0.2/229$$
$$\mathrm{Chl a }\left(\mathrm{mg}/\mathrm{g}\right)=\mathrm{Chl a }+\mathrm{Chl b}$$

where Chl a is chlorophyll a content; Chl b is chlorophyll b content; Car is carotenoid content; and Chl is the total chlorophyll content.

The net photosynthetic rate (Pn), intercellular CO2 concentration (Ci), transpiration rate (Tr), stomatal conductance (Gs), and chlorophyll fluorescence dynamic parameters were monitored using a Li-6400 portable photosynthesis system (LI-COR, Nebraska, USA). Photosynthesis-related experiments were conducted from 08:30 to 11:00 on clear days using the flag leaf (counting from the apex) of each plant. Three individual plants were used in each experiment.

The WUE of YX and YX-yl was calculated using the following formula:

$$\mathrm{WUE}=\mathrm{Pn}/\mathrm{Tr}$$

Chloroplast microstructure

Fresh leaves of YX and YX-yl at the fifth-leaf stage were collected, cut into 1 × 0.5 cm2 sections, and fixed overnight in a solution of 2.5% glutaraldehyde. The samples were washed with 0.1 M phosphate-buffered saline (pH = 7.4) three times for 15 min and then post-fixed in 1% osmium tetroxide for 2 h. The fixed leaf samples were washed with 0.1 M phosphate-buffered saline (pH = 7.4) three times for 15 min and then dehydrated in gradient ethanol solutions twice for 20 min. The dehydrated tissue was immersed in a 1:1 mixture of embedding agent-acetone for 1 h and then immersed in a 3:1 mixture of embedding agent-acetone for 3 h. Subsequently, the tissues were embedded in resin. Ultrathin sections were made from the embedded leaf samples using a Leica UC7 Ultra-wave slicer (Leica UC7; Leica, Japan), stained with uranium lead, and examined using a JEM 1230 transmission electron microscope (Jeol, Tokyo, Japan).

Content of precursor substances for chlorophyll synthesis

The flag leaves of YX and YX-yl at the seedling stage were collected to analyze the content of precursor substances. Each sample contained three biological replications. ALA, PBG, Proto IX, Mg-Proto IX, and Pchlide levels were measured using a UV-1800 ultraviolet/visible spectrophotometer (UV-1800pc; MAPADA). Proto IX, Mg-Proto IX, and Pchlide content was calculated using the following formula:

$${\mathrm{C}}_{\mathrm{Mg}-\mathrm{Proto IX}}\left(\mathrm{ng}/\mathrm{g}\right)= {0.06077\mathrm{A}}_{590}-{0.01937\mathrm{A}}_{575}-{0.003423\mathrm{A}}_{628}$$
$${\mathrm{C}}_{\mathrm{Mg}-\mathrm{Proto IX}}\left(\mathrm{ng}/\mathrm{g}\right)= {0.1806\mathrm{A}}_{575}-{0.04036\mathrm{A}}_{628}-{0.04515\mathrm{A}}_{590}$$
$${\mathrm{C}}_{\mathrm{Pchlide}}\left(\mathrm{ng}/\mathrm{g}\right)= {0.03563\mathrm{A}}_{575}-{0.007225\mathrm{A}}_{590}-{0.02955\mathrm{A}}_{575}$$

RNA extraction, library construction, and RNA-seq

Total RNA was extracted from YX and YX-yl leaf samples using a TRIzol RNA isolation kit (Invitrogen, USA). The purity, concentration, and integrity of the extracted RNA were assessed using a Nanodrop spectrophotometer (ND-1000; Name-drops Technologies, Wilmington, DE, USA), a Qubit 2.0 Fluorometer (Thermo Fisher Scientific, USA), and an Agilent 2100 bioanalyzer (Agilent Technologies, Santa Clara, CA, USA).

A sequencing library was prepared using the Illumina TruSeq RNA Sample PreKit. Briefly, mRNA was purified from the total RNA using poly T oligo-attached magnetic beads, fragmented, and reverse-transcribed into cDNA. Adapters were then ligated onto the cDNA molecules, and the fragments were amplified by PCR. Sequencing was performed on paired-end reads (2 × 150 bp) using an Illumina HiSeq 2500 sequencing platform. mRNA sequencing and analysis were conducted by Biomarker Biotechnology Co., Ltd. (Bei**g, China).

Raw data were processed to remove low-quality reads and adapters. Next, clean data were assembled using the Trinity software (Broad Institute, Cambridge, MA, USA) [44] to obtain the unigene library of broomcorn millet. The unigene library was mapped with NR, Swiss-Prot, GO, COG, KOG, eggNOG4.5, and KEGG databases using the BLAST software [45] to obtain the annotation information of unigenes. Next, functions of unigenes were obtained using the HMMER software [46] to map unigenes with the Pfam database.

Gene expression analysis and detection of differentially expressed genes

Clean reads were mapped with the Unigene library using the Bowtie2 software [47] and then combined with RSEM [48] for evaluation of gene expression levels. The expression level of unigenes is represented by the FPKM value [49].

Gene expression analysis was performed using the DESeq2 software [50] (www.bioconductor.org). Genes with false discovery rate (FDR) < 0.01 and fold change (FC) ≥ 2 were considered as DEGs between YX and YX-yl. GOseq enrichment analysis of DEGs was performed using the GOseq R package. Enrichment analysis of DEGs with annotation information between YX and YX-yl was conducted using the TopGO software (Version 2.18.0). Significantly enriched GO terms in DEGs were defined as GO terms with a corrected p-value of ≤ 0.05.

The enrichment of DEGs in KEGG pathways was assessed using the KOBAS software [51, 52], and significantly enriched KEGG pathways were defined as those with FDR ≤ 0.05.

Validation of DEGs by qRT-PCR

To validate the sequencing data, we first selected nine DEGs for qRT-PCR. Total RNA was extracted from the leaf tissues of YX and YX-yl. The PrimeScript™ RT reagent Kit with gDNA Erase (Takara, Dalian, China) was used to synthesize first-strand cDNA. qRT-PCR was performed using gene-specific primers in a total volume of 10 μL comprising 5 μL of SYBR Premix Ex Taq II, 3 μL of ddH2O, 1 μL of primer mix (1:1 mix of forward and reverse primers at 10 μmol/μL each), and 1 μL of cDNA as a template. The reaction conditions were as follows: 30 s at 95 °C, followed by 39 cycles of 5 s at 95 °C, 30 s at the annealing temperature, and 5 s at 60 °C to 95 °C. The primers used for qRT-PCR analysis are listed in Table S6. The actin gene was used as an internal standard. Relative expression levels were calculated as 2−△△ct [53].