Introduction

Leaf colour mutations such as albinism, light green, yellowish, stripe fleck, and others have been identified in many plants1. Using mutant plants, researchers have shown that leaf colour changes may alter chloroplast ultrastructure, leaf physiological and biochemical characteristics, and genetic mechanisms regulating leaf colour. These changes usually lead to reduced chlorophyll (Chls a and b) content in leaf colour mutants. Physiologically, Chl plays an important role in harvesting light and transferring energy. It is involved in chloroplast development, photomorphogenesis, and signal transduction between the nucleus and chloroplast. Thus, leaf colour plays important roles in regulation of photosynthetic efficiency.

Many explanations for leaf colour variations have been provided. One reason for low Chl content in many higher plants is Chl biosynthetic deficiency2,3,4,5. Studies on the Chl biosynthetic process have focused on Arabidopsis thaliana and rice. Previous reviews and studies involving these two species revealed that at least 27 genes controlling 15 steps were involved in Chl biosynthesis from glutamyl-tRNA to Chls a and b6. Other reasons for Chl deficiency include deficient signal transduction between the nucleus and chloroplast3,7, restrained heme feedback8,9, impaired synthesis and importation of chloroplast proteins10,11,12,13,14,15, and harmful photooxidation16,17,18,19. Overall, the molecular mechanisms regulating leaf colour phenotypes are very complex.

Cytological, physiological, and proteomic studies on leaf colour mutations have been reported for Brassica napus20,21,22,23,24,25,26. A genetic analysis of a leaf colour mutation, Cr, in B. napus was first reported in a Chl-deficient mutant with a yellowish-leaf phenotype controlled by one recessive gene. This yellowish mutant had been introduced into male sterile lines and used as a marker to produce hybrid seed26. Another Chl-deficient mutant, BnaC.ygl, was controlled by one recessive gene mapped onto B. napus chromosome C06. The marker linked to it was within 0.03 cM27. A spontaneous yellowish-leaf mutant of B. napus, BnChd1, was controlled by two pairs of recessive genes. Its gene (or locus), BnChd1-1, was inferred to be Bra040517 located on an A01 chromosomal region based on functional and expression analyses of genes within the mapping region28.

We herein report the inheritance of a dominant Chl-deficient mutant (cde1) isolated from an ethyl methanesulfonate (EMS)-mutagenized B. napus line, as well as the gene mapping, together with a saturated single nucleotide polymorphism (SNP) linkage map. Our findings offer new insights into the dominant yellowish-leaf phenotype.

Results

Chl content

Seedlings of the cde1 mutant have yellowish leaves that gradually become green at the budding stage (Fig. 1a–c). The buds of the cde1 mutant are light-yellow, whereas those of wild-type plants are green (Fig. 1d). The Chl content of cde1 mutants and wild-type plants are provided in Table 1. The mutants had significantly lower Chls a and b and total Chl content, and a lower Chl a/b ratio than the wild-type plants.

Figure 1: Morphological performance of the cde1 mutant and its wild-type.
figure 1

(a) Wild-type plant (left) and cde1 mutant plant (right) at the 6-week-old stage. (b) Leaves of the Wild-type (left) and cde1 mutant plant (right). (c) The leaves of the cde1 mutant plant (left) and Wild-type plant (right) after the budding stage. (d) The buds of the cde1 mutant plant (left) and Wild-type plant (right).

Full size image
Table 1 Chl content in the leaves of cde1 mutants and wild-types in the F2 population in Brassica napus.
Full size table

The influence of leaf colour on yield traits

The relationship of leaf colour with agronomic traits was investigated by randomly sampling the mature plants from Chl-deficient and green seedlings in the F2:3 populations. The results revealed that the 1,000-seed weight and yield of individual Chl-deficient plants were lower than that of normal green plants although many of the other traits were not significantly different between the two groups (Table 2). This clearly indicated that the yellowish leaves at the seedling stage reduced the yields.

Table 2 Difference in agronomic traits between the yellow and green-leaf seedlings in the F2:3 populations of Brassica napus.
Full size table

Inheritance of the yellowish-leaf trait

A cross between the canola variety ‘zhongshuang11’ (‘ZS11’) and our mutant cde1 generated F1 plants with the same yellowish leaves as the mutant. By selfing the F1 plants, we generated 157 F2 plants that consisted of 119 yellowish-leaf and 38 green-leaf plants. Results of a chi-squared test indicated that the segregation pattern agreed with the Mendelian segregation ratio of 3:1 (yellowish-leaf vs. green-leaf plants) (Table 3). Thus, the yellowish-leaf trait is controlled by a single dominant gene. Additionally, results for the F2:3 family populations from the four selfed heterozygous plants also agreed with an expected Mendelian inheritance ratio of 3:1 (yellowish-leaf vs. green-leaf plants) (Table 3). This further demonstrated that the yellowish-leaf trait was controlled by one dominant gene. The dominant inheritance of this Chl deficiency phenotype is rare in nature and differs from the inheritance reported for most previously described yellowish-leaf mutants23,24,26,27,28.

Table 3 Segregation of the F2 generation and its four F2:3 families of Brassica napus.
Full size table

Construction of a high density SNP map

From the (‘ZS11’ × cde1) F2 population, 89 plants were used for SNP genotyping. Although the chip has 52,157 SNP markers, the 14,377 polymorphic markers were used to construct the linkage map after removing invalid markers. The 19 linkage groups (LGs) produced by the JoinMap 4 software contained 2,878 bins representing 13,347 SNP markers (Supplementary Data). The total length of the map was 1,968.6 cM. The longest LG was C03 at 173.9 cM and the shortest was C02 at only 41.8 cM. The mean interval between adjacent markers was 0.69 cM (Table 4). These data indicated that our linkage map based on SNP markers was likely sufficient to map the locus controlling the yellowish-leaf trait.

Table 4 Basic information for 19 linkage groups constructed with the (‘ZS11’ × cde1) F2 population of Brassica napus.
Full size table

Mapping the CDE1 locus

By linking the leaf colour phenotypes with the corresponding SNP marker data for the F2 population, the yellowish-leaf trait locus was mapped onto LG C08, positioned in a 0.9 cM interval between SNP markers M29912 and M29878. The CDE1 locus co-segregates with the bin M29880, which has eight co-segregated SNP markers: M29910, M29890, M29888, M29887, M29884, M29883, M29882, and M29881. By searching databases containing the genomes of B. oleracea and B. napus cv. ‘ZS11’, we found that probe sequences for the nine co-segregating SNPs completely matched genome sequences. In the B. napus cv. ‘ZS11’ genome (http://www.ncbi.nlm.nih.gov/assembly/GCA_000686985.1), the interval harboring the CDE1 locus was 378.32 kb (Fig. 2).

Figure 2: Mapping interval for CDE1 on LG C08 and its corresponding physical map of chromosome C08 of Brassica napus cv. ‘ZS11’.
figure 2

This linkage map was a part of LG C08 in supplemental table S1, and the physical map was obtained by inquiries of ‘ZS11’ genome database with SNP probes.

Full size image

To narrow the mapping interval for further study, simple sequence repeat (SSR) markers uniformly covering the interval were developed based on the target interval genomic sequence. In total, 135 pairs of SSR primers were developed (Supplementary Table S1). These markers were used in the mapping population. However, only two pairs of primers, BnC08Y56 and BnC08Y66, were polymorphic and acted as co-dominant markers. These two SSR markers generated the expected products (Fig. 3) and co-segregated with the CDE1 locus in the F2 population.

Figure 3: PCR products of SSR codominant markers BnC08Y56 and BnC08Y66 in the F2 population.
figure 3

Here, ‘y*’ and ‘g’ indicates products from homozygous yellowish-leaf plants and green-leaf plants, respectively. ‘y’ indicates a heterozygous yellow-leaf plant. (a,b) are the PCR products of SSR markers BnC08Y56 and BnC08Y66, respectively.

Full size image

We conducted additional experiments using 620 individuals of the F2:3 populations generated from the cross involving the tightly linked polymorphic markers BnC08Y66 and BnC08Y56. We obtained four recombinants between the CDE1 locus and two SSR markers with green-leaf phenotype. The recombinants identified by this SSR marker experiment were confirmed by a SNP marker experiment that adopted the same technology as that used for the F2 plant genotyping. Furthermore, the SNP marker experiment demonstrated that of the four recombinants, two occurred between M29910 and M29890 and two occurred between M29888 and M29887. These data may help estimate the recombination rate. The distance between CDE1 and SNP marker M29890 was 0.24 cM, and the distance between CDE1 and SSR marker BnC08Y66 was 0.42 cM (Fig. 4). Therefore, the locus was on a 311-kb chromosomal interval between SNP markers M29912 and M29890 of the B. napus cv. ‘ZS11’ genome (http://www.ncbi.nlm.nih.gov/assembly/GCA_000686985.1).

Figure 4
figure 4

Fine mapping of CDE1 on chromosome C08 of Brassica napus with SSR and SNP markers in the F2:3 populations.

Full size image

Brassica segments homologous to the mapped interval

An analysis of the homologous segments of the interval carrying the CDE1 locus was necessary. The searches in Brassica genome databases including that of B. rapa (http://brassicadb.org/brad/), B. oleracea (http://brassicadb.org/brad/), B. napus cv. ‘Darmor-bzh’ (http://www.Genos cope.cns.fr/ brassicanapus/), and B. napus cv. ‘ZS11’ (http://www.ncbi.nlm.nih.gov/assembly/GCA_000686985.1) with the aid of SNP probes generated five homologous segments (Table 5).

Table 5 Genes on the mapped segments and their homologous Brassica segments.
Full size table

Alignments among the homologous segments with dot matrix software (http://blast.ncbi.nlm.nih.gov/Blast.cgi), except that of B. napus cv. ‘Darmor-bzh’ that contains a large gap of approximately 70 kb, showed that the two C08 segments were structurally similar, with slight differences in the structure of the two A09 segments (Supplementary Fig. S1). However, the similarities between the two A09 segments were almost the same as those between the two C08 segments. These results were consistent with the evolutionary relatedness among Brassica species.

The genes of B. rapa, B. oleracea, and B. napus cv. ‘Darmor-bzh’ have been annotated. However, the target mapping interval of the genome of B. napus cv. ‘Darmor-bzh’ has genomic gap of 70 kb in length because of its imperfect genome assembly, leading to the absence of part of the gene annotation. Unfortunately, the genes in the genome of B. napus cv. ‘ZS11’ (one parent in our study) have not been annotated. This made it difficult to find candidate genes responsible for the yellowish-leaf trait. Nevertheless, the extensive information available regarding the Brassica genome annotation might be helpful. By use of the homologue search approach, we annotated the genes for the un-annotated segment, and replenished gene information for the annotated segments as shown in Table 5. Based on these works, the mapped B. napus cv. ‘ZS11’ segment was discovered to contain 22 genes, and the other homologous segments contained 17–23 genes (Table 5). Additionally, the gap in the C08 segment from ‘Darmor-bzh’ may be filled with a 77-kb segment from the unanchored “chrC08_random” pseudo-molecule, based on its homology to the other four segments (Supplementary Fig. S2). As a result, the C08 segment from ‘Darmor-bzh’ was discovered to contain 23 genes.

Candidate genes analysis

Based on the gene codes on the homologous segments, as illustrated in Table 5, we inquired the genome databases of Brassica genus (http://brassicadb.org/brad/, http://www.genoscope.cns.fr/brassicanapus/, and http://www.ncbi.nlm.nih.gov/assembly/GCA_000686985.1) to find gene function information based on the putative function of their homologous genes in A. thaliana (Table 5). BnaC08g34740D, BnaC08g34800D, and BnaC08g34840D are directly involved in Chl biosynthesis. The gene BnaC08g34740D homologous to AT2G22880, encoding VQ motif-containing protein (VQ12), acting as a cofactor of WRKY transcription factors, may be the candidate gene responsible for the yellowish-leaf trait because the mutant vq8-1 in Arabidopsis exhibits the yellowish leaf phenotype, whereas VQ8 is homologous to VQ12 and has an N-terminal signal peptide that is predicted to be chloroplast targeting29. BnaC08g34800D is homologous to AT1G78490 that encodes cytochrome P450 protein CYP708A3 with heme-binding activity, and may participate in Chl metabolism. BnaC08g34840D is homologous to AT5G4258, which encodes a cytochrome P450 protein CYP705A12 and has a similar function as CYP708A3. Thus, two genes, BnaC08g34800D and BnaC08g34840D may regulate leaf colour because of functional changes30. All the three genes may lead to recessive Chl-deficient phenotype, but they are far from being good candidate genes for cde1 mutant to explain the dominance mechanism. Our DNA sequencing experiments showed that these three genes have no sequence differences between the cde1 mutant and ‘ZS11’.

BnaC08g49180D and BnaC08g34850D are homologous to AT2G22760 and AT2G22750, respectively, both of which belong to the family of bHLH transcription factor genes. A member of the bHLH transcription factor family, PIF3 (Phytochrome-Interacting Factor 3), is a direct phytochrome reaction partner in the photoreceptor’s signalling network31. Besides PIF3, bHLH transcription factor family members PIF1, PIF4, and PIF5 are also negative phytochrome-interacting protein and can negatively regulates Chl biosynthesis32,33,34. Because phytochrome drives the assemblage of chloroplast and enhancement of Chl content35, we infer that BnaC08g49180D and BnaC08g34850D are good candidate genes for the yellowish-leaf trait because they might control the reaction of phytochrome and alter Chl content. DNA sequencing experiments revealed that there is no sequence difference for BnaC08g49180D and one amino acid residual difference at the site +97 for BnaC08g34850D between the cde1 mutant and ‘ZS11’, thus more focus need to be put on BnaC08g34850D in future to explore its relationship with the mutation of cde1.

In the A09 segment of B. napus cv. ‘Darmor-bzh’, BnaA09g42360D may be related to Chl biosynthesis because it is homologous to AT1G60550, which encodes DHNS (DHNA synthase) that is involved in phylloquinone biosynthesis and research has confirmed that it participated in Chl biosynthesis30. As BnaA09g42360D has no homologous genes in the other four homologous segments, it was not deemed a candidate gene of the CDE1 locus in the present study.

Four genes (BnaC08g34760D, BnaC08g34770D, BnaC08g34780D, and BnaC08g49190D) in the CDE1 mapping interval of the C08 segment of B. napus cv. ‘Darmor-bzh’ encode proteins of unknown function, which therefore could not be excluded because their relationship with the yellowish-leaf mutant is unknown to date. Furthermore, other genes with functionalities such as transcription factor or RNA splicing, etc, in the mapping interval, cannot also be excluded from candidate gene list before the gene responsible for the yellowish leaf phenotype is demonstrated.

Discussion

Seedlings of the cde1 mutant isolated from an EMS-mutagenized line exhibit a yellowish-leaf phenotype and Chl deficiency. At the budding stage, cde1 leaves gradually turn green. The leaf colour trait of cde1 is controlled by one dominant gene, rather than recessive genes, such as Cr, BnChd1, and BnaC.ygl26,27,28 in Brassica napus. To date, research has showed that most leaf colour mutations are caused by recessive mutated genes, whereas few are controlled by dominant genes. Because leaf colour cannot be restored to green in a heterozygote, it might be concluded that the locus controlling leaf colour does not consist of genes involved in biosynthesis from glutamyl-tRNA to Chls a and b6. In our work, BnaC08g34850D which belongs to bHLH transcription factor may be regarded as candidates for dominant CDE1 locus. Previous reports revealed that PIFs as members of bHLH transcription factor family act as negative regulators of Chl biosynthesis by directly binding to G-box motifs in the PORC (protochlorophyllide oxidoreductase) and FeChII (ferrochelatase) promoters in Chl biosynthetic pathway, may lead to yellowish leaf phenotype36. Our former iTRAQ-based quantitative proteomics analysis of the cde1 mutant and its corresponding wild-type has demonstrated that the gene expressions in Chl biosynthesis and photosynthesis pathways20 were down-regulated in the yellowish leaves. These results are consistent with theoretical prediction analysis about the candidates. Thus BnaC08g34850D in the mapping interval may be probably the important candidate gene for the dominant yellowish locus. However, the unknown protein genes and other genes functioning as transcription factor or RNA splicing, etc, within the mapping interval, cannot be excluded.

The use of SNP markers has benefited plant genotyping efforts because of the numerous distinct markers and high genome coverage37,38,39,40,41. Whether obtained from chip hybridization or DNA sequencing experiments, SNP marker data have been commonly used for gene or locus identification and molecular breeding. The Brassica 60 K SNP BeadChip Array has helped to advance rapeseed research efforts42,43,44. It has recently enabled the efficient construction of several high-quality saturated linkage maps over a short period45,46,47,48,49. Our current study also produced a saturated B. napus map, with 2,878 bins containing 13,347 SNP markers and a total length of 1,968.6 cM. We used 89 rapeseed plants from a segregating population for SNP genotyping and linkage mapping. We then developed SSR markers for fine-mapping of the target locus. We pre-mapped the locus using SNP markers and fine-mapped the locus with SSR markers based on the genome sequence. We adopted this mapping strategy by linking SSR markers with SNP markers. This represents a good mapping strategy because of the associated reduced SNP genotyping costs and convenience of using SSR or other markers. Additionally, designed SNP polymorphic markers may have low coverage in the mapping interval.

Our findings provided important information regarding the gene responsible for the yellowish-leaf mutation based on the mapped interval and genetic information. However, the mapped interval length of 311 kb is likely too long to enable the identification of specific genes for subsequent gene function evaluations. In fact, because of the lack of polymorphic SNP or SSR markers, the mapping interval obtained is probably the shortest possible interval for the mapping population used. To further shorten the mapped interval, one strategy could be to use parents with polymorphic markers to construct a new inheritance population for mapping. Another strategy could involve resequencing of the mutant accession to confirm the sequence differences that regulate the phenotypic variation in leaf colour.

B. napus originated approximately 7,500 years ago by hybridization between B. rapa and B. oleracea, followed by chromosome doubling50. B. rapa and B. oleracea, as parents of B. napus, are highly related evolutionarily. Based on this knowledge, it can be concluded that short segments, such as that listed herein (less than 312 kb), are usually conserved structurally among species and contain a parallel set of genes. Therefore, comparisons among closely related homologous segments are beneficial in that they allow replenishment of annotation information for species with incomplete information in the genome database. The comparisons we conducted generated gene annotation information and allowed gap filling. As such, our results on the gene annotation based on the alignments are reliable and necessary for finding the candidate genes. In our work, we found that the homologous segments on A09 and C08 chromosomes contain nearly the same set of genes, but the homologous segments on A09 chromosomes contain smaller gene sets than those on the C08 segments. This was likely the result of genome (A and C) evolution in the Brassica genus.

Methods

Plant materials

The B. napus Chl-deficient mutant, cde1, was originally isolated from EMS-mutagenized B. napus line NJ7982 at Nanjing Agricultural University in China. The yellowish-leaf cde1 mutant was crossed with the canola variety ‘ZS11’, then selfed to produce the F2 generation. We phenotyped the F2 population plants at the 6-week-old stage, and DNA from 13 wild-type plants and 76 mutants, and DNA from the recessive parent ‘ZS11’, was selected for SNP genotyping. In total, 38 wild-type and 26 mutant plants of the F2 population at the 20-week-old stage were prepared for Chl content determination. The F2 population and its four F2:3 family populations were used for genetic segregation analysis. We randomly selected 30 cde1 mutant plants and 30 wild-type plants from the F2:3 populations for analysis of agronomic traits. The 90 individuals for SNP genotyping and 620 individuals from the F2:3 families were scanned for SSR markers. All materials were grown in the fields of the Jiangpu Experimental Station at Nanjing Agricultural University. Plants were sown uniformly in a row of 2.5 m length with 15 individuals in each row and the rows were spaced at a distance of 0.4 m.

Chl content determination

Chl was extracted from 0.2-g samples of fresh leaves using 50 mL of 80% acetone (LingFeng, Shanghai, China). Absorbances of total Chl, Chl a, and Chl b were determined using an established procedure20.

Construction of a SNP map

Total DNA was extracted from fresh leaves using a modified cetyl-trimethylammonium bromide (CTAB) method51. The DNA samples were diluted to 200 ng uL−1 and then genotyped using the Brassica 60 K SNP BeadChip Array. The DNA sample preparation, hybridization to the BeadChip, and imaging of the arrays were completed by the Beijing Emei Tongde Development Co. Ltd. (Beijing, China). Allele-calling for each locus was performed using the GenomeStudio genotyping software, v2011.1 (Illumina, Inc.). Cluster definitions were based on genotype data from the rapeseed individuals. The SNP markers were named using “M” plus the index numbers assigned by GenomeStudio.

The polymorphic SNP markers were first sorted into different bins. The first marker within each bin was selected as the representative of the bin and used to construct the LG with JoinMap 4 software52. The SNP markers were first grouped at the linkage logarithm of odds 13.5. The marker order and distances in each LG were calculated using JoinMap 4, with the mapping function of Kosambi and the Regression mapping algorithm.

Mapping of the CDE1 locus

Based on the SNP map, the CDE1 locus was determined and potential physical regions containing the CDE1 locus were identified by aligning SNP probe sequences and co-segregated SNP markers in the genomes of B. napus and B. oleracea (one ancestral parent of B. napus). The interval sequence was downloaded (http://www.genoscope.cns.fr/brassicanapus/data/ and http://brassicadb.org/brad/downloadOverview.php) for bioinformatics analysis and SSR primer development to confirm the interval detected for CDE1.

Based on the sequences of the physical regions of B. napus containing the CDE1 locus, 1,621 SSR loci were identified using SSRHunter 1.3 software53 with a 6-bp motif maximum and three repeat minimum. A total of 135 SSR loci with a 150 bp sequence on both sides were selected to design primers using the Primer Premier 5.0 software54. The PCR conditions were as follows: denaturation at 95 °C for 10 min, followed by 35 cycles of 95 °C for 30 s, annealing for 40 s, and 72 °C for 40 s and a final extension step at 72 °C for 10 min. Two SSR markers, BnC08Y56 (223 bp expected size; forward primer: 5′-TTTAACCGGGACTTGAGA-3′; reverse primer: 5′-TTGGGCTAATGAACCTTT-3′; 49.8 °C annealing temperature) and BnC08Y66 (249 bp expected size; forward primer: 5′-GAGGAGCGACAAGATGAA-3′; reverse primer: 5′-TAAGTACCACCGAAAGCA-3′; 49.2 °C annealing temperature), were polymorphic and used to genotype the population and fine-map the CDE1 locus. The recombinant individuals identified by SSR marker scanning were further genotyped by the Brassica 60 K SNP BeadChip Array.

Brassica segments homologous to the mapped interval and candidate gene analysis

Whole genome sequences of B. rapa, B. oleracea, B. napus cv. ‘Darmor-bzh’, and B. napus cv. ‘ZS11’ were downloaded from a public database (http://brassicadb.org/brad/, http://www.genoscope.cns.fr/brassicanapus/, and http://www.ncbi.nlm.nih.gov/assembly/GCA_000686985.1)50,55,56. The alignment of sequences containing the CDE1 locus to genomic sequences was performed using BLASTN (http://blast.ncbi.nlm.nih.gov/). Segments covering all adjoining similar sequences with an E value ≤ 1e-50 were considered homologous segments in the corresponding genome. The similarity of these homologous segments was assessed by dot matrix analysis. A bioinformatics analysis of annotated genes in these homologous regions was then completed.

Additional Information

How to cite this article: Wang, Y. et al. Fine mapping of a dominant gene conferring chlorophyll-deficiency in Brassica napus. Sci. Rep. 6, 31419; doi: 10.1038/srep31419 (2016).