Introduction

Chlorophyll (Chl) is a green pigment in higher plants and other photosynthetic organisms and plays an essential role in photosynthesis by capturing light energy and converting it into chemical energy1. There are two kinds of Chl in higher plants, Chl a and Chl b. Chl a and Chl b have the same chemical structure except in the side chain at C7 of the tetrapyrrole, where Chl a has a methyl group and Chl b has a formyl group. Chl a take part in the reaction of both the photosynthetic reaction centers and the light-harvesting antennae, whereas Chl b only occurs in the antenna complexes and can control the photosynthetic antenna size to enable optimal utilization of available light2,3,4,5. Increasing Chl content is an important approach for improving crop yield by increasing the photosynthetic rate and promoting photoassimilate accumulation, which is also a focus of breeders6,7,8.

Chl is synthesized in 16 steps from glutamyl tRNA in higher plants, and Chl b is synthesized by chlorophyllide a oxygenase (CAO) from Chl a2,9. Recombinant CAO can catalyze the reaction from chlorophyllide a to chlorophyllide b by a two-step oxidation reaction3. CAO is considered as the only enzyme responsible for Chl b synthesis, and defective CAO results in Chl b content cut down in Chlamydomonas reinhardtii2 and Arabidopsis thaliana10,11. The sequence of CAO is highly conserved from cyanobacteria to higher plants12,13,14,15. CAO consists of three domains: A, B, and C16. The A domain, i.e., the N-terminal conserved sequence, is considered to perceive the existence of Chl b and participates in regulating CAO protein levels17,18. The C domain includes a [2Fe–2S] Rieske center and mononuclear iron motif that exercise its catalytic activity12. The B domain, between the A and C domains, is less conserved, which may be related to the protein stability of CAO19,20,21.

In higher plants, the coding region of most genes (80–85%) is composed of exons and introns22. An exon is the part of DNA that is retained after pre-RNA is cut or modified. It appears in the gene sequence of mature RNA and can be translated to protein to perform gene functions. Introns in DNA are transcribed into the pre-RNA, but during RNA splicing, the introns in the transcripts of primary genes are removed, and the exons are connected to produce mature mRNA. Although introns cannot be transcribed and translated into proteins, many studies have found that introns play an important role in alternative splicing and regulation of gene expression23,24,25, especially when the intron is placed close to the 5′ end of the gene26,27. He et al.28 first showed that the length of intron 1 in BrMYB2 regulates anthocyanin biosynthesis and controls the development of a purple or white head in Chinese cabbage. Lasin et al.29 showed that the two introns of ATSUC1 are required for gene expression in roots.

Although some leaf color-related genes have been cloned in Chinese cabbage, there is still a difference in the mechanism and characteristics; thus, more leaf color mutant genes need to be cloned. This study completed a mapping of the pale green mutant gene (Brpgm) in Chinese cabbage and revealed that a single nucleotide polymorphism exists in BraA10g007770.3C (BrCAO), resulting in premature translational termination. Based on the identification of phenotypic, genetic, and photosynthetic characteristics of the pale green mutant (pgm), it is suggested that BrCAO plays a critical role in Chl synthesis and yield in Chinese cabbage.

Results

Mutant characteristics identification

The pgm plants appeared evenly pale green and showed normal growth (Fig. 1a). In the vegetative growth and reproductive growth stages, pgm and FT were both able to develop normal heading, flowering, and seed setting. Biomass was analyzed by measuring the fresh weight and dry weight of the whole plant at the age of 6 weeks. No significant differences were observed both on fresh weight and dry weight of pgm and FT (Fig. 1b).

Figure 1
figure 1

Phenotypic characterization and ultrastructure of chloroplasts of pale green mutant (pgm) and wild-type (FT). (a) Phenotypic characterization of pgm (right) and FT (left) at 6 weeks. (b) The fresh weight and dry weight in pgm and FT at 6 weeks. (c, d) Ultrastructure of chloroplasts in FT (c) and pgm (d). Scale is shown at the bottom.

Full size image

Photosynthesis pigment content and chloroplast structure

Compared with FT, pgm showed a more significant reduction in both total Chl and Car content (Table 1). In pgm, Chl a, Chl b, and Car, 50.63%, 82.46%, and 59.65% reduction caused a higher level of Chl a/b and a stable level of Car/Chl, suggesting that the obstruction of the Chl b synthesis pathway was greater than that of the Chl a synthesis pathway.

Table 1 Content of photosynthetic pigments of leaves in pale green mutant (pgm) and wild-type (FT).
Full size table

To investigate whether the chloroplast development of pgm affected, we observed the ultrastructure of chloroplasts in pgm and FT using transmission electron microscopy (Fig. 1c,d). The chloroplasts of pgm were plump and oval in shape, but the thylakoid lamellae were loose, with more starch grain accumulation. The FT chloroplasts were elongated and elliptic, with well-developed thylakoid membrane systems, large number of grana, tight and orderly arrangements, abundant matrix lamellae, and more stacking.

Photosynthetic parameters and fluorescence kinetic parameters

The photosynthetic index of FT and pgm was measured, and it was found that the Pn of pgm was 6.97% lower than that of FT; however, the decrease was not significant (Table 2). The Gs and Ts of pgm did not change significantly, but the Ci of pgm increased significantly.

Table 2 Photosynthetic parameters and fluorescence kinetic parameters of leaves in pale green mutant (pgm) and wild-type (FT).
Full size table

Compared to that of FT, the F0, Fm, primary photochemical efficiency of PSII (Fv/Fm), and NPQ of pgm were significantly reduced. However, the actual photochemical efficiency of PSII (ΦPSII) and PQ of pgm were significantly increased (Table 2); the lack of significant change in the Pn in pgm was possibly attributable to this.

Inherited characteristic

We constructed genetic populations by FT and pgm to figure out the inheritance pattern of pgm (Table 3). The inbred offspring of FT (P1) and pgm (P2) showed green and pale green, respectively, proving both mutant and wild-type could be inherited stably. F1 (P1 × P2) and rF1 (P2 × P1) plants showed green color, indicating that the pale green character of pgm showed nuclear inheritance and was controlled by recessive nuclear genes. In the F2 population, there was character segregation, and the ratio of green plants to pale green plants was 3:1, indicating that the mutation was controlled by one pair of recessive nuclear genes. BC1 (F1 × P1) was green, while in BC1 (F1 × P2), the green plants were separated from the pale green plants by 1:1, which further confirmed the single recessive inheritance of the mutant trait.

Table 3 Genetic analysis of pgm in Chinese cabbage.
Full size table

The Brpgm located on chromosome A10 via BSR-seq

Based on the genetic analysis of pgm, we constructed a large-scale F2 segregated population of pgm and K23. The two RNA mixing pools (GP-pool and PGP-pool) were separately constructed by green phenotype and pale green phenotype plants of F2 population and sequenced by Illumina for BSR-Seq analysis. Finally, 11,625,396 and 9,526,538 raw reads and 11,558,596 and 9,477,028 clean reads with an average length of 148.51 bp and 148.34 bp, respectively, were obtained (Supplementary Table S1). The clean data were compared with the reference genome, and 80% of the reads of PGP and GP were mapped. Single nucleotide polymorphisms (SNPs) were detected in the GP pool and PGP pool (Supplementary Table S2). SNPs with a coverage depth greater than 3X were screened in the two samples simultaneously, and the Euclidean Distance (ED) of these SNPs was calculated. The ED values of the SNP loci were processed by the fifth power method, ED5, to avoid the influence of background noise on the experimental results. A total of 48,827 SNPs were applied to map the ED5 values, depending on the distribution of each SNP locus on the chromosome. A statistical peak of ED5 was found on chromosome A10, with the top 1% as the threshold, and four correlation intervals were determined (A10: 0.12–2.50 Mb, 3.17–4.49 Mb, 8.56–9.73 Mb, and 10.31–12.51 Mb) (Fig. 2 and Supplementary Table S3).

Figure 2
figure 2

The distribution of ED5 values on chromosomes.

Full size image

Based on the above information, SSR primers were designed in four target regions, and the polymorphism of primers was screened by mutant pgm and K23 for further experiments (Supplementary Fig. S1a). Primer sequence information is detailed in Supplementary Table S4. We identified 30 individuals with a pale green phenotype in the F2 population as a small group to validate BSR-Seq results. The results showed that SSR11-45 and SSR1-10 on chromosome A10 were linked to Brpgm and located on both sides of Brpgm, according to linkage analysis (Supplementary Fig. S1b,c). These results illustrated that the Brpgm gene was mapped on chromosome A10 between SSR11-45 and SSR1-10.

Mapping of the Brpgm to a 4499.6 Kb interval based on linkage analysis

Using 2184 F2 individuals with a pale green phenotype, we developed new primers and polymorphic markers that were screened and amplified, and calculated the recombination rate. Finally, the markers SSR10-17, SSR12-9, SSR9-27, SSR7-18, SSR5-1, and SSR3-1 were closely linked to the target gene. SSR3-1, SSR5-1, SSR7-18, and SSR1-10 were on the same side, while SSR10-17, SSR12-9, SSR9-27, and SSR11-45 were on the other side of the target gene. The genetic distances between the mutant gene and SSR 7–18 and SSR 9–27 were 0.11 and 0.02 cM (Fig. 3b), respectively. The mutation gene Brpgm is located in a physical range of 4499.6 Kb (Fig. 3c), in which there are 336 genes (Supplementary Table S5).

Figure 3
figure 3

Genetic and physical maps of the Brpgm gene locus in Brassica rapa. (a) The mutant gene Brpgm was initially located between the molecular markers SSR10-17 and SSR5-1 in the A10 linkage group, with genetic distances of 2.20 and 0.43 cM, respectively. (b) The mutation gene Brpgm was located between SS9-27 and SSR 7–18 with genetic distances of 0.02 and 0.11 cM, respectively. (c) The physical size of the candidate region was 4499.6 Kb, containing 336 candidate genes.

Full size image

Candidate gene prediction by whole-genome resequencing

Limited by the size of the mapping population and the number of recombined individuals, the candidate region could not be further narrowed by linkage analysis. Therefore, whole-genome resequencing was carried out in FT and pgm lines to detect the mutation sites. Sequencing results showed that just one homozygous non-synonymous SNP was located in the candidate region (Table 4), which consistent with the EMS mutagenesis principle (variations in G-to-A and C-to-T)30 and located on BraA10g007770.3C. The gene annotation indicated that BraA10g007770.3C (BrCAO) encodes chlorophyllide a oxygenase (CAO), which converts chlorophyllide a to chlorophyllide b by catalyzing two successive hydroxylations at the 7-methyl group of chlorophyllide a. Clone sequencing revealed that the 180th nucleotide of full length was changed from G to A, and the 88th nucleotide of CDS was deleted (Fig. 4a,b). Based on the gene information (Fig. 4c and Supplementary Table S6), we found that the SNP was located in the last position of the first intron, which changed the splicing mode and resulted in the deletion of the first nucleotide in the second exon. Sequencing of six recombinant individuals of the two most recent markers verified that the SNP was co-isolated with the mutant phenotype (Fig. 4a,b). The deletion of G in CDS caused a frameshift mutation, leading to the early termination of protein translation (Fig. 4d).

Table 4 Single nucleotide polymorphism (SNP) in the candidate region between wild-type (FT) and pale green mutant (pgm).
Full size table
Figure 4
figure 4

Sequence alignment and gene structure. (a) DNA sequence alignment of wild-type (FT), pale green mutant (pgm), and F2 recombinant individuals: 204, 332, 354, 919, 1153, and 1542. (b) cDNA sequence alignment of FT, pgm, and F2 recombinant individuals. (c) Gene structure of BrCAO. (d) BrCAO protein sequence of FT and pgm. The red box indicates the mutation position.

Full size image

Expression analysis and enzyme assay

The expression pattern of BrCAO in leaves was determined using qRT-PCR. The results showed that the expression of BrCAO in pgm was upregulated in four stages (cotyledon, seedling, rosette, and heading stages), notably in the seedling stage (Fig. 5a). In the seedling stage, the activity of CAO was determined by ELISA and the result showed that the BrCAO activity was significantly lower in pgm than in FT (Fig. 5b).

Figure 5
figure 5

Expression analysis of BrCAO and enzyme assay of BrCAO. (a) The expression pattern of BrCAO in leaves of four stages (cotyledon, seedling, rosette, and heading stage) in pale green mutant (pgm) and wild-type (FT). (b) The CAO activity in pgm and FT. U/L = activity unit. The ‘*’ represents significant differences at 5% levels.

Full size image

Discussion

Leaf color mutants are usually related to Chl content, including the regulatory network of genes involved in Chl biosynthesis, degradation, and chloroplast development. In this study, we isolated pgm, which showed a pale green phenotype and normal growth (Fig. 1). Based on BSR-Seq, we developed two SSR markers, SSR9-27 and SSR7-18, which are closely linked to Brpgm on chromosome A10 (Figs. 2 and 3). Although the genetic distance between SSR9-27 and SSR7-18 was minimal, there was a large physical distance between the two markers, containing 336 candidate genes. After many rounds of expanding this group, we used 2184 recessive individuals to map the cause gene, but a low recombination rate appeared in the localization interval. The location interval (A10: 3,654,958–8,154,544) was located in the centromere interval (A10: 5,449,476–8,087,520) of the A10 chromosome31. Previous studies have found that it is difficult to map by centromere intervals because of the high content of repetitive sequences and low recombination frequency32,33,34,35,36. Therefore, we think that the larger and more difficult-to-narrow physical distance of the Brpgm candidate region is related to its location near the centromere.

Whole-genome resequencing further detected the SNP in the candidate region between FT and pgm lines and screened only one homozygous non-synonymous SNP located on BraA10G007770.3C (BrCAO). Parents and F2 recombinant individual sequencing verified that this SNP co-separated with the pale green phenotype. BrCAO encodes chlorophyllide a oxygenase (CAO), which is the only essential enzymatic step for Chl b formation3. Tanaka et al.2 and Espineda et al.10 characterized ATCAO mutants in Arabidopsis thaliana, which have reduced levels of Chl b and higher AtCAO mRNA levels. Lee et al.4 identified Line 1C-039-43, whose the first intron of OsCAO1 existed a insertion, which was deficient in Chl b, producing pale green leaves. In our study, pgm had lower chlorophyll content with a higher Cha/b value and lower BrCAO activity with higher BrCAO mRNA levels. Mueller et al.15 by multiple independent alleles validated the candidate gene of a light green chlorina phenotype (due to the reduction in Chl b) as HvCAO. Pattanayak et al.37 revealed that overexpression of CAO in tobacco could increase chlorophyll (Chl) b biosynthesis and decrease the Chl a/b ratio.

RNA splicing is essential for the development and evolution of organisms, and the key to splicing recognition depends on the determination of the splicing sites. The splicing sites of pre-mRNA in eukaryotic cells mostly follow the GU-AG rule; that is, the base groups of intron 5′ end (donor site) and 3′ terminal (acceptor site) are almost GU and AG22,38. In this study, the 180th nucleotide of BrCAO full length varied from G to A in pgm, which is located in the last position of the first intron. Although this SNP is located in the intron, it leads to the deletion of the 88th nucleotide in CDS, located in the first nucleotide of the second exon. This result may be associated with RNA splicing. The 180th nucleotide of full length changing from G to A, resulting in the last two positions of the first intron (RNA splice sites) changed from AG to AA. The first nucleotide (G) of the second exon and the last nucleotide (A) of the first intron forms a new RNA splice site AG, which results in the deletion of the first nucleotide (G) of the second exon in CDS. The deletion of nucleotide G in CDS resulted in a frameshift mutation, leading to the early termination of protein translation.

As the main product of Chinese cabbage plants, the leaves have always been of concern. In this study, pgm appeared to be stably pale green and possessed a lower Chl content. Unlike most yellow leaf mutants, the deficiency of Chl content in pgm did not lead to weak growth and weight reduction but regulated the color of leaves. This result may be related to the fact that Pn was not significantly different and due to higher PQ values with lower NPQ values. In pgm, most of the light energy absorbed by the PSII antenna pigment is used for photochemical reaction electron transfer. Only a small part cannot be used for photosynthetic electron transfer but is dissipated in the form of heat. Despite the decrease in chlorophyll content, the net photosynthetic efficiency was stable and showed a younger color in pgm. Therefore, pgm could be considered a new germplasm for leaf color improvement without affecting yield.

In conclusion, we identified pgm created by EMS mutagenesis in Chinese cabbage. BSR-Seq and linkage analysis mapped Brpgm in the centromere interval of chromosome A10. Whole-genome resequencing analysis predicted that the target gene was BrCAO. The mutation of BrCAO in pgm occurred at the splice site of the first intron, which resulted in the early termination of the protein translated by BrCAO. pgm showed a pale green phenotype and possessed lower Chl content with a higher Chl a/b ratio, imperfect chloroplast structure, and lower NPQ. Pn and biomass of pgm was not significantly altered, which suggested that pgm may be used for color improvement in Chinese cabbage breeding.

Materials and methods

Plant materials

The wild-type (FT), doubled haploid (DH) line of Chinese cabbage, was used as the wild type and test materials in ethyl methane sulfonate (EMS) mutagenesis. The pgm was identified in the mutant populations. The Pak choi inbred line, “K23” with deep green leaves was used to construct the segregating population with pgm. All materials were grown and provided by Shenyang Agricultural University, Shenyang, China.

Pigment content measurement

Chl and carotenoid (Car) content were determined using a DU 800 UV/Vis Spectrophotometer (Beckman Coulter, La Brea, CA, USA) according to the method outlined by Arnon39, with some modifications. The fifth leaves of 6-week-old plant were harvested and submerged in 80% acetone under dark conditions for 24 h. The extracts were measured at 663, 645, and 470 nm. Pigment concentrations were calculated as described by Holm40. Three plants were measured per treatment, and each sample carried out three times repeats.

Determination of photosynthetic parameters

At the age of 6 weeks, the fifth leaves were selected to determine the photosynthetic parameters using a portable photosynthetic system (CIRAS-2, PP Systems, USA). Measurements were recorded for three individual plants per treatment at a sunny day. Data was automatically recorded until a steady net photosynthetic rate (Pn) was attained. The photosynthetic parameters consisted of Pn, stomatal conductance (Gs), intercellular CO2 concentration (Ci), and transpiration rate (E).

Analysis of Chl fluorescence kinetics

At approximately six-weeks-old, FT and pgm plants with the same growth were selected to measure fluorescence parameters using a Chl fluorescence imaging system (IMAGING-PAM, Walz, Germany), which is a platform instrument of the facility of Horticulture at the College of Horticulture. On a sunny morning, the plants were kept away from light for 20 min. Then, the fifth true leaf of the plants were removed and placed in the instrument. The pulse intensity was set to 4500 μmol m−2 s−1, the pulse time was set to 0.8 s. Then the related Chl fluorescence kinetics parameters were determined and recorded. Three biological repeats were identified in each material.

Transmission electron microscopy analysis

The same leaf parts of the 6-week-old plants were cut into 2 × 6 mm pieces, pre-fixed in 3% (w/v) glutaraldehyde and stored overnight at 4 °C. After rinse with 1% phosphoric acid buffer, the samples were fixed with 1% osmium acid for 2 h. The following procedures were carried out as described by Zhao et al.41.

Genetic analysis

FT (P1) and pgm (P2) lines were crossed and produced F1 (P1 × P2) and rF1 (P2 × P1). F1 plants were self-pollinated to produce an F2 population. The BC1 populations were derived by the backcrosses of FT and pgm lines, respectively, with F1. The phenotype characterization and segregation ratio of each generation (P1, P2, F1, rF1, F2, and BC1) were recorded and analyzed using the χ2 test. All measurements in the experiments were analyzed using a random design.

Bulked segregant RNA-sequencing (BSR-seq)

To map the mutant gene Brpgm of the pgm line, we used K23 and the pgm line to construct the F2 population. In the F2 mapping population, 50 green phenotype plants and 50 pale green phenotype plants with the same growth were selected to extract total RNA. The green phenotype RNA mixing pool (GP-Pool) and pale green phenotype RNA mixing pool (PGP-Pool) were constructed by mixing green phenotype and pale green phenotype plant samples, respectively. RNA was extracted using a plant total RNA extraction kit (Tiangen, Beijing, China), following the manufacturer’s procedure. BSR-Seq was performed and analyzed as described by Zhao et al.42.

DNA extraction and PCR amplification

Total genomic DNA was extracted from fresh leaves using the CTAB method according Murray and Thompson43, with minor modifications. A 10 μl system was used as the PCR amplification reaction system, and the PCR procedure was as follows: pre-denaturation at 95 °C for 5 min, one cycle; denaturation at 95 °C for 30 s; renaturation at 56 °C for 30 s; extension at 72 °C for 1 min, 35 cycles; final extension at 72 °C for 5 min.

SSR marker analysis, linkage analysis, and genetic map construction

Based on the location interval of BSR-Seq, the genome sequence information was downloaded from the Brassica database (http://brassicadb.org/brad/datasets/pub/Genomes/Brassica_rapa/V3.0/). Primer Premier 5.0 software (Premier Inc., Charlotte, USA) was used to design primers. Polyacrylamide gel electrophoresis was used to screen the primers with polymorphisms between the parents. Then, linkage analysis was performed using the polymorphic markers in F2 individuals with a pale green phenotype. Segregation data were used to construct a linkage map of the F2 population using Join Map 4.044. The genetic map distances (cM) were calculated according to the method of Kosambi45.

Whole-genome resequencing

The genomic DNA of FT and pgm were extracted using DNA Secure Plant Kit (Tiangen, Beijing, China) for whole-genome resequencing. A DNA library with 400 bp of insert size was constructed and sequenced using next-generation sequencing and Illumina HiSeq paired-end sequencing (Illumina, San Diego, USA). The raw data were analyzed after removal of joint contamination, quality filtering, and length filtering to generate clean data. The BWAMEM program was used to map the filtered clean data to the reference genome (http://brassicadb.org/brad/datasets/pub/Genomes/Brassica_rapa/V3.0/). GATK software46 and ANNOVAR software47 were used to extract SNPs (single nucleotide polymorphism) and annotate SNPs, respectively.

Clone sequencing

The full-length and CDS of BrCAO were amplified using FL-BrCAO primers and CDS-BrCAO primers, respectively (Supplementary Table S7). PCR products were purified and ligated to the pGEM-T Easy Vector (Promega, USA). The vectors were transformed into competent E. coli cells. After culturing, plasmids were extracted and sequenced using GENEWIZ (Suzhou, China). Sequencing data were analyzed using DNAMAN V6 software (Lynnon BioSoft, Canada).

Total RNA extraction and gene expressive analysis

Total RNA samples were extracted from fresh leaves of different stages (cotyledon, seedling, rosette, and heading stages) using a plant total RNA extraction kit (Tiangen, Beijing, China). cDNA was synthesized using FastQuant RT Super Mix 13 (Tiangen, Beijing, China) Quantitative real-time PCR (qRT-PCR) amplification was carried out in QuantStudio 6 (Life Technologies, California, USA) using SYBR Green PCR Master Mix (Takara Bio Inc., Kusatsu, Japan) in a 20 μl reaction mixture. Gene-specific primers were designed using Primer Premier 5.0, and the ACTIN gene was used as the internal control (Supplementary Table S8). The qRT-PCR amplification reaction system and procedure was described as Huang et al.48.

Enzyme activity assay

An enzyme-linked immunosorbent assay (ELISA) kit (Meimian Industrial Co., Ltd., Jiangsu, China) was used in accordance with the manufacturer’s instructions to determine the activity of CAO. Leaves of six-week-old plants (0.2 g FW) were homogenized in phosphate buffered saline (PH7.4). The supernatant was obtained by centrifugation at 12,000×g and used for the analysis. The experimental process was conducted according to the manufacturer’s instructions for ELISA kits.

Statistical analysis

The t-test was used to analyze the significant differences at a significance level of 0.05.

Ethical approval

The study was performed in accordance with relevant guidelines and regulations.