Deciphering variations, identification of marker–trait associations and candidate genes for seed oil content under terminal heat stress in Indian mustard (Brassica juncea L. Czern & Coss) germplasm stock

Abstract

This research paper investigates the variability in seed oil content (SOC) in Indian mustard (Brassica juncea L.) under terminal heat stress (THS) conditions. A genetic stock of 488 genotypes of B. juncea was evaluated over two years and grouped into five classes based on the reduction in oil content under THS compared to normal sown crop. Based on heat susceptibility index (HSI), a diverse panel of 96 genotypes was selected and evaluated under THS. Twenty-two heat-tolerant donor genotypes were identified, including introgression lines derived from B. tournefortii, B. carinata and Erucastrum cardaminoides. This study is the first to report on marker–trait associations for SOC in B. juncea under THS using a GWAS approach. Furthermore, candidate genes associated with abiotic stress tolerance and lipid metabolism were identified near the significant SNPs, emphasizing their role in SOC regulation under stress. Notable candidate genes include BjuA003240 (encoding for alcohol-forming fatty acyl-CoA reductase), BjuA003242 (involving in lipid biosynthesis), BjuA003244 (associated with mitochondrial functions and stress tolerance), and BjuA003245 (related to MYB transcription factors regulating lipid biosynthesis). This study provides valuable insights into the genetic basis of SOC variation under THS in B. juncea, highlighting potential breeding targets for improved heat stress resilience in Indian mustard cultivation.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13205-024-03985-w.

Introduction

In Indian mustard (Brassica juncea L. Czern & Coss), the economic value is determined by oil yield, which is influenced by seed yield (SY) and seed oil content (SOC). This crop is mainly prevalent in the Indian subcontinent and has expanded to arid, low-rainfall areas due to its adaptability to limited-moisture conditions (Parker 1999). Additionally, B. juncea exhibits resistance to blackleg disease, a serious threat to B. napus, a major canola oil crop (Woods et al. 1991). It also possesses superior heat and drought tolerance, as well as resistance to seed shattering. These unique characteristics have renewed interest in B. juncea as a climate-resilient crop. The allopolyploid B. juncea (AABB; 2n = 36) has emerged through hybridization events between B. rapa (AA, 2n = 20) and B. nigra (BB, 2n = 16) in various geographies, such as the Mediterranean, Irano-Turanian and Saharo-Sindian regions (Nagaharu 1935).

Indian mustard crop is well adapted to low moisture cropping system due to its low water requirement (80–240 mm). However, various abiotic stresses, such as high temperatures at germination and terminal stages, drought, salinity and cold spells, lead to low productivity of this crop (Wang et al. 2003). Nearly, 30.7% area (1.81 mha) of Indian mustard is under rainfed farming in India and the problem of heat stress particularly during reproductive phase have become major concern due to change in climatic conditions such as shortening of winter duration as well as delayed sowing (Sandhu et al. 2021). This study evaluates effect of terminal heat stress (THS) on SOC due to delayed sowing as well as by simulating the THS condition under controlled poly-tunnel environment.

Terminal heat stress (THS) specifically refers to heat stress during the reproductive phase, which is critical as it can result in flower abortion and significant reductions in both seed yield (Hall 1992), seed oil yield (Singh and Bhajan 2016) and seed quality (Ahmad et al. 2021). In India, delayed sowing of mustard crops after harvesting rice and cotton exposes the crop to high-temperature stress (Chauhan et al. 2009), leading to reduced productivity and a significant gap between potential and actual yield. The stage of the crop plays a crucial role in determining the extent of yield loss due to heat stress, with THS causing more economic losses compared to the seedling stage, as the crop has less time for recovery in the latter stage. Delayed planting also shortens the vegetative phase, advances flowering time and reduces the seed development period (Srivastava and Balkrishna 2003; Dhaliwal et al. 2007). The quantitative inheritance of oil content, coupled with high genotype-environmental interaction, results in low selection efficiency for this trait. Heat stress during the post-anthesis (seed-filling) stage negatively affects the movement of photosynthates to developing sinks, thereby impacting seed and oil yield in oilseed Brassicas.

To establish thermo-tolerance in Indian mustard, it is crucial to assess the variability of this trait under heat stress conditions. Genome-wide association study (GWAS) and quantitative trait locus (QTL) mapping are commonly used to understand the genetic architecture of complex quantitative traits at the whole genome level. GWAS provides a higher resolution compared to traditional bi-parental linkage mapping and is not limited to specific populations. Genotyping by sequencing (GBS) can be used to discover thousands of markers across the genome of interest in a population within a few weeks (Davey et al. 2011). While seed oil content is a complex quantitative trait, its genetic and molecular mechanisms are not yet fully understood (Xiao et al. 2019). There are only a few studies that have explored oil content through GWAS.

A fixed diverse germplasm stock (derived through continuous selfing/sibbing) of 488 Indian mustard genotypes comprising landraces; released Indian varieties/cultivars; accessions collected from different countries and developed introgression lines, is available with Punjab Agricultural University (PAU), Ludhiana. The germplasm collection was evaluated for oil content for two years at PAU- Regional Research Station, Abohar, Fazilka district of Punjab state, India under two planting regimes. This location is situated in the arid zone of Punjab which experiences higher temperature than rest of Punjab during onset of summer season. This region accompanied by delayed planting of germplasm was explored for natural screening of germplasm for terminal heat stress. From this stock, a panel of 96 genotypes, which represented differential response to SOC, was used to determine heat susceptibility index (HSI). From 96 genotypes, a set of 71 diverse genotypes was used to constitute association mapping panel for GWAS and candidate gene identification for SOC.

Materials and methods

Constitution of diversity panel for GWAS

A fixed diverse germplasm stock of 488 genotypes, comprising introgression lines, land races, old cultivars and advance breeding lines, was used to explore the variability for seed oil content (SOC) under different planting regimes. The germplasm collection was planted at PAU Regional Research Station, Abohar, Fazilka (represented by semi-arid plains with scattered sand dunes; located at 30.15° N, 74.19° E) for two years during 2015–16 and 2016–17 under two environments. Two environments represented two sowing regimes: set I sown on November 14 (represented normal sown conditions; abbreviated as NS); set II sown on December 2 (represented late sown conditions to expose the crop to terminal heat stress; abbreviated as THS). The test material was planted in an alpha lattice design in two replications with each genotype sown in a plot of two rows of two-meter length at a row to row and plant to plant spacing of 30 cm and of 10 cm, respectively. At maturity, the crop was harvested manually. Siliques were taken from ten random plants, dried properly and thrashed. Five gm dried intact seed sample of each genotype was taken and SOC was determined using non-destructive bench top nuclear magnetic resonance (NMR) spectrometer (MQC23, Oxford Instruments, UK) (ISO 10565:1998). Standard Brassica seed samples with known oil contents (estimated using Soxhlet apparatus) were used to calibrate the instrument. Percent oil content reduction (OCr%) for each genotype, due to heat stress during the reproductive phase of crop, was computed using formula:

$$\mathrm{OCr}\%=\frac{{\text{OC}}_{\text{NS}}-{\text{OC}}_{\text{THS}}}{{\text{OC}}_{\text{NS}}}\times 100$$

OC_NS: oil content under NS condition, OC_THS: seed yield under THS condition

The heat susceptibility index (HSI) for oil content (%) for each genotype was calculated using formula given by Fischer and Maurer (1978).

$$\text{HSI}=\frac{1-\frac{{\text{OC}}_{\text{THS}}}{{\text{OC}}_{\text{NS}}}}{1-\frac{\overline{{\text{OC}}_{\text{THS}}}}{\overline{{\text{OC}}_{\text{NS}}}}}$$

$\overline{{\mathit{OC}}_{\mathit{NS}}}$: Mean oil content under NS; $\overline{{\mathit{OC}}_{\mathit{THS}}}$: mean oil content under THS.

Phenotyping and genotyping of the diversity panel

From stock of 488 genotypes, the diversity panel of 96 genotypes, representing variable HSI for oil content and differential response to OCr, was planted at the research farm of Punjab Agricultural University, Ludhiana (30.90° N, 75.86°E) during the year 2017–18 under two different environments, created by two planting systems. In the environment I, the panel was sown under natural field conditions (NS) and in environment II, the same set of panel was sown in the same field but covered with poly-tunnels during the reproductive phase i.e., the onset of flowering to seed ripening stage of the crop, to artificially create terminal heat stress for the crop. Poly-tunnels were shaped by placing an iron rod frame in field and covering the crop by polythene sheets (3 mm thickness) from 10.00 am to 3.00 pm every day. In each environment, the panel set was sown in an alpha lattice design with two replications; with each genotype in 3-m row length. Row to row distance of 30 cm and plant to plant distance of 10 cm was maintained. Weather data on temperature was recorded under normal and heat stress conditions (Table 1). All agronomic and protection measures were taken to raise a healthy crop. Data were recorded at physiological maturity for SOC and analyzed statistically.

Table 1.

Weekly maximum and minimum temperature during reproductive phase of B. juncea in two environments

Week	Growth stage of crop	Normal sown conditions (NS)		Terminal heat stress conditions (THS)
		Minimum temperature (°C)	Maximum temperature (°C)	Minimum temperature (°C)	Maximum temperature (°C)
II (8–14 Feb)	Siliquae initiation	6.4	21.6	17.9	30.6
III (15–21 Feb)	Siliquae development	12.9	25.8	22.5	34.5
IV (22–28 Feb)	Siliquae development	8.4	24.5	20.9	36.4
I (1–7 March)	Seed development and filling	9.7	25.2	21.5	37.7
II (8–14 March)*	Seed filling	9.4	19.4	21.2	33.5
III (15–21 March)	Seed filling and onset of physiological maturity	11.6	26.2	22.3	38.8

^*32.2 mm rainfall observed during this week

Statistical analysis of phenotypic data

Phenotypic distributions of oil content were plotted as histograms with fitting normal curve. Analysis of variance (ANOVA) and adjusted means (LSMEANS) for alpha lattice design were calculated using Proc GLM procedure in SAS 9.3 (SAS Institute 37) for each environment. Mean broad-sense heritability (h²_bs) was calculated by variance components estimated in Generalized Linear Model (GLM) using the formula:

$$h_{\text{bs}}^2=\frac{{\sigma }^{2}g}{{\sigma }^{2}g+\frac{{\sigma }^{2}gy}{k}+\frac{{\sigma }^{2}e}{\mathit{rk}}}$$

where ${\sigma }^{2}g$ is the genotypic variance, ${\sigma }^{2}gy$ is the variance due to genotype and environment interaction, ${\sigma }^{2}e$ is the error variance, k is the number of environments and r is the number of replications.

DNA extraction and GBS data

From 96 genotypes, a set of 71 diverse genotypes (listed in Table S2) was used to constitute association mapping panel for GWAS and candidate gene identification for SOC. High-quality genomic DNA was isolated from young tissue of genotypes using CTAB (cetyltrimethyl ammonium bromide) method given by Doyle et al. (1990). Genotyping by sequencing (GBS) was outsourced from AgriGenome Labs Pvt. Ltd., India using double digestion Restriction site-Associated DNA sequencing (ddRAD-seq) platform. Quality clean GBS reads were aligned with the reference genome of B. juncea (www.brassicadb.cn) using Bowtie2 (version 2–2.2.9) program with default parameters (Langmead and Salzberg 2012). Samtools version 0.1.18 (Li et al. 2009) was used for identification of population SNPs followed by filtration with missing data > 20% and MAF < 5%.

Genome-wide association analyses

A total of 18,258 filtered SNPs were used as genotypic data for GWAS of SOC, using FarmCPU (Fixed and Random Model Circulating Probability Unification) (Liu et al. 2016) package in R statistical software (R Core Team 2018). FarmCPU is a multi-locus model approach and addresses the confounding problems of mixed linear models (MLM) using both fixed effect model and the random effect model iteratively (Liu et al. 2016). GWAS was performed for each environment (THS and NS) separately, to identify associated SNPs in specific environments. The multiple associated markers are estimated by incorporating kinship values in the random effect model (Liu et al. 2016). Population structure was estimated with principal component analysis (PCA) of genotypic data using GAPIT version 3.0 (Genomic Association and Prediction Integrated Tool) package (Lipka et al. 2012). Principal components (PCs) and kinship computed from genotypic SNP data were incorporated in the FarmCPU model. The significantly associated SNPs based on their associated p-values were identified as putative genomic regions associated with SOC in B. juncea.

Identification of putative candidate genes

LD analysis was performed across the genome and a cut off of r² = 0.2 was used for determining LD decay. Flanking region of 50 kb on either side of significant SNPs was explored for putative candidate genes. Annotation of genes present in the confidence interval of SNPs was obtained from Brassica database (http://brassicadb.cn/#/FlankingRegion/). Functions of the predicted candidate genes were reviewed to establish their importance for improving SOC and THS tolerance in B. juncea.

Results

Documenting variation for seed oil content in germplasm stock of B. juncea under terminal heat stress

Increasing SOC is one of the most important targets for B. juncea breeding. This study documented the variation for oil content in a fixed diversity stock of 488 Indian mustard genotypes under NS as well as natural THS conditions. The frequency distribution of SOC in test genotypes was presented as histogram with befitting normal curve (Fig. 1a, 1b; Table S1). HSI values have revealed wide variability in the stock displaying most susceptible (HSI = 6.77) to most tolerant (HSI = −4.57) genotypes (Fig. 2a; Table S1). The OCr% under THS in comparison to optimum conditions ranged from −3.16 to 4.68 (Fig. 2b, Table S1). Significant reductions in oil content (positive OCr%) under THS condition and positive HSI values verified the adverse effect of heat stress on the growth and development of B. juncea. However, many genotypes have negative OCr% revealing increase in OC under heat stress and thereby indicates the tolerance of these genotypes to high-temperature stress.

Fig. 1.

Distribution of seed oil content (%) in 488 genotypes of B. juncea under different planting conditions: a Terminal heat stress, b Normal sown. The histograms display the frequency distribution of seed oil content across genotypes, providing insights into the variation observed under both terminal heat stress and normal sowing conditions

Fig. 2.

Distribution of 488 genotypes of B. juncea for a Oil content-based heat susceptibility index (HSI) and b Percent oil content reduction (%OCr). The histograms illustrate the variation in HSI and %OCr among the genotypes, highlighting the range of responses to terminal heat stress

Top 22 genotypes (out of 488) with negative HSI and OCr% were identified as potential donors for better performance under heat stress conditions and hence, suggested for utilization in commercial breeding after evaluation in larger plots (Table 2). Amongst these 22 identified donors, six were introgression lines viz., PTJ-3–72, PTJ-3–79 and PTJ-3–102 derived from Brassica tournefortii x B. juncea; DAR-1 and MSC-5 from B. juncea x B. carinata hybridization, and MCP-12–211 from B. juncea x Erucastrum cardaminoides. The differential response of 96 genotypes to THS based on HSI and OCr%, was further validated by evaluating under controlled conditions. A significant increase in weekly temperature regimes during different reproductive stages of the crop under controlled conditions in comparison to NS conditions (control) authenticated the screening conditions of panel for THS using poly-tunnels (Table 1). Significant variation was observed in genotypes for SOC under both NS and THS environments (Table 3). Under THS conditions, OC ranged from 35.25 to 41.76 with mean value 38.64 ± 0.14, while under NS conditions, it ranged from 36.40 to 42.58 with mean value 39.96 ± 0.14. The pooled analysis of variance revealed significant G x E interaction for SOC, inferring the effect of THS on this trait. This also explains low (52%) broad-sense heritability on pooled basis. Significant moderate correlations (r = 0.43) were obtained between NS and THS environment for SOC. Panel evaluation under controlled environment depicted variability for OCr% under THS- and SOC-based HSI (Table S3). Out of 96 genotypes, 71 representing set was used for GWAS studies. The distribution pattern of SOC (%) of 71 genotypes under THS and NS was presented in Fig. 3a and b, respectively.

Table 2.

Heat Susceptibility Index (HSI) and percent oil content reduction (OCr%) of top tolerant 22 potential donor genotypes of B. juncea to terminal heat stress under field conditions

Genotype name	Germplasm type	HSI	OCr%
PTJ-3–72	B. juncea introgression line from B. tournefortii x B. juncea	−4.57	−3.16
CM-20–2	Low erucic acid line	−4.54	−3.14
RC-214	Advance breeding line	−4.53	−3.13
RRN-624	Advance breeding line	−4.47	−3.09
GMCN-139	Germplasm line	−4.44	−3.07
PCR-3	Advance breeding line	−4.44	−3.07
GLM 1–2	Quality germplasm line	−4.40	−3.04
PTJ-3–102	B. juncea introgression line from B. tournefortii x B. juncea	−4.40	−3.04
PUSA BOLD DT	Determinate plant type version of adapted cultivar	−4.39	−3.03
PBG-1007	Advance breeding line	−4.38	−3.03
PTJ-3–79	B. juncea introgression line from B. tournefortii x B. juncea	−4.37	−3.02
GMCN-100	Germplasm line	−4.37	−3.02
ISB-89	Germplasm line	−4.35	−3.01
CCJJ-1	Germplasm line	−4.34	−3.00
CSR-103	Germplasm line	−4.34	−3.00
DAR-1	Introgression B. juncea line from B. juncea x B. carinata hybridization	−4.34	−3.00
CRL-1359–18-21	Quality breeding line	−4.33	−2.99
CSR-171	Germplasm line	−4.32	−2.99
JMWR 946–1-13	White rust resistant line	−4.32	−2.99
MCP-12–211	Derived B. juncea line from B. juncea x Erucastrum cardaminoides	−4.31	–2.98
RMM-09–4	Advance breeding line	−4.29	−2.97
MSC-5	Derived line from B. juncea x B. carinata	−4.29	−2.96

Table 3.

Mean, standard error, range, mean squares and coefficient of variation (CV) of oil content in panel of 96 B. juncea genotypes under terminal heat stress and normal environments

Environments	Mean ± SE	Range	Mean square (genotype)	Mean square (G x E)	Coefficient of variance	Broad sense heritability (%)
Terminal Heat Stress	38.64 ± 0.14	35.25–41.76	3.13***	–	0.49	98.0
Normal Sown	39.96 ± 0.14	36.40–42.58	2.89***	–	1.36	81.0
Pooled	39.29 ± 0.12	36.85–41.89	4.22***	2.03***	0.95	52.0

^***Significance at p < 0.001

Fig. 3.

Distribution of seed oil content (%) in association mapping panel of 71 genotypes of B. juncea under different planting conditions: a Terminal heat stress, b Normal sown. The histograms depict the distribution of seed oil content across the association mapping panel, allowing comparison between terminal heat stress and normal sowing conditions

Genotyping by sequencing (GBS) data analysis

A range of 1.7–6.1 million raw reads were obtained through GBS of diverse panel of 71 genotypes of B. juncea. The GC percent varied from 39.8% to 44.5%, while the Q30 value ranged from 90.75% to 94.29%. Overall, between 72.23% and 82.42% of the reads aligned to the reference genome across all genotypes. After SNP calling using samtools version 0.1.18, a total of 30,673 SNPs were obtained. Filtering the data resulted in 18,258 SNPs with less than 20% missing data and a minor allele frequency (MAF) greater than or equal to 5%. These SNPs were uniformly distributed across the genome of B. juncea, with the highest number on the B08 chromosome and the lowest on the A10 chromosome (Fig. S1; Table S4). The linkage disequilibrium (LD) decay in the population was estimated to be 50 kb, where the r2 value dropped below 0.2 (Fig. S2).

Principal component analyses of B. juncea diverse panel

Principal components (PCs) were derived from SNP marker data to capture population structure variation in Brassica juncea. PC1 and PC2 elucidated 5.1% and 4.1% of the total genetic variance, respectively. The scatterplot depicting PC1 against PC2 revealed three broad clusters within the population structure, indicating the presence of subgroups (Fig. S3). Consequently, first three principal components (PCs) were incorporated as a covariate in the GWAS model as fixed effect and kinship among the genotypes was incorporated as a random effect to account for the relatedness among individuals.

GWAS for seed oil content

GWAS for seed oil content was done with FarmCPU which is a multi-locus model that differs from the widely used mixed linear model (MLM) as it takes care of overwhelmed false positives and painful false negatives by dividing modified MLM into fixed and random effects and used them iteratively (Liu et al. 2016). It addresses the confounding problem of mixed linear models (MLM) using population structure and multiple associated markers in fixed effect model. Under THS and NS, total seven significant marker–trait associations were identified (Table 4; Fig. 4). Under THS, three SNPs viz. A01.2143072, A06.26228835 and B04.17559973 located on chromosomes A01, A06 and B04 respectively, have significant association with SOC. In NS environment, four SNPs viz. A02.22908480, A04.17533560, B04.34761164 and B07.21583568 located on chromosomes, viz. A02, A04, B04 and B07 were identified as significant associations with SOC.

Table 4.

Significant SNP markers associated with oil content (OC) in 71 B. juncea genotypes under terminal heat stress (THS) and normal sown (NS) conditions

Traits	SNPs associated	Chromosome	Position on genome (bp)	Alleles	MAF	Effect	−log10(p) value	Previous reports
THS	A01.2143072	A01	2,143,072	G/T	0.18	−0.61	5.32
	A06.26228835	A06	26,228,835	A/T	0.06	0.86	5.31
	B04.17559973	B04	17,559,973	A/G	0.42	−0.50	6.69
NS	A02.22908480	A02	22,908,480	C/G	0.45	−0.42	5.92	[3]
	A04.17533560	A04	17,533,560	A/C	0.19	0.61	6.49
	B04.34761164	B04	34,761,164	C/T	0.06	−0.66	6.04	[1], [2], [3]
	B07.21583568	B07	21,583,568	C/T	0.08	−0.74	6.56	[2], [3]

[1] Mahmood et al. (2006), [2] Yadava et al (2012), [3] Rout et al. (2018)

Fig. 4.

Manhattan and quantile–quantile plots generated from genome-wide association analysis results for oil content: a Normal sown—NS, b Terminal heat stress—THS. *Chromosome 1–10 represents “A” genome mentioned as A01–A10 and chromosome 11–18 represents “B” genome mentioned as B01 to B08 in manuscript

Identification of candidate genes

Based on LD decay, 50 kb flanking region on either side of the seven significantly associated SNPs with SOC were studied and 52 genes were obtained in the vicinity of six SNPs viz., three SNPs under THS and NS each have 31 and 21 genes, respectively (Table 5). No gene annotation was found for SNP “A02.22908480”. Under THS, 17 genes were studied for SNP “A01.2143072”, out of which, 16 genes (BjuA003240 to BjuA003255) were found within 50 kb region around the SNP. These genes were involved in abiotic stress tolerance, regulation of lipid biosynthesis, RNA–DNA hybrid ribonuclease activity, fatty acid reductase and proteolysis. Gene BjuA003238 was found ~ 55 kb upstream of the SNP, encoding for peroxidase enzyme, which are involved in biosynthesis of phenylpropanoids.

Table 5.

Protein family and gene(s) in the confidence interval of significant SNPs associated with seed oil content

Environment	SNP position chromosome (bp)	Confidence interval (50 kb up/downstream)	Genes in confidence interval (gene id)	Protein family (description)
Terminal heat stress (THS)	A01 (2,143,072)	2,093,072–2193072	BjuA003238	Phenylpropanoid biosynthesis (KO ID: K00430)
			BjuA003240	AITR3; DIG-LIKE 2; DIL2 (AT5G40790)
			BjuA003241	Nucleic acid binding and RNA–DNA hybrid ribonuclease activity
			BjuA003242	Fatty acid reductase 3
			BjuA003243	Phosphatidylinositol signaling system
			BjuA003244	OKI1; OKINA KUKI (AT4G33760)
			BjuA003245	myb-like protein X;(source: Araport11)
			BjuA003246	No Result
			BjuA003247	hypothetical protein
			BjuA003248	proteolysis; cysteine-type peptidase activity
			BjuA003249	ATCAPE1; CAP-DERIVED PEPTIDE 1
			BjuA003250	ATCAPE3
			BjuA003251	CBS domain protein (DUF21); (source: Araport11)
			BjuA003252	G patch domain protein;(source: Araport11)
			BjuA003253	ABERRANT GROWTH AND DEATH 2; AGD2; ARF-GAP DOMAIN 2
			BjuA003254	Cysteine-rich TM module stress tolerance
			BjuA003255	dynamin 1-like protein [EC:3.6.5.5]
	A06 (26,228,835)	26,178,835–26,278,835	BjuA000608	Cellulase (glycosyl hydrolase family 5) protein;(source: Araport11)
			BjuA024420	alpha-Linolenic acid metabolism
			BjuA024421	AP2/ERF domain
			BjuA024422	RNA–DNA hybrid ribonuclease activity
			BjuA024423	Protein dimerization activity
	B04 17,559,973)	17,509,973–17,609,973	BjuB028757	UNC-50 (ARF exchange factor)
			BjuB028758	GH3 auxin-responsive promoter
			BjuB028759	DUF760 (Protein of unknown function)
			BjuB028760	Gibberellin regulated protein
			BjuB028761	DUF1206 (Protein of unknown function)
			BjuB028762	Mpv17/PMP22 (22-kDa peroxisomal membrane protein)
			BjuB028763	Mpv17/PMP22 (22-kDa peroxisomal membrane protein)
			BjuB028764	Transcriptional regulator of RNA polII, SAGA, subunit
			BjuB028765	Reverse transcriptase-like
Normal sown (NS)	A02 (22,908,480)	22,858,480–22958480	Not available	Not available
	A04 (17,533,560)	17,483,560–17583560	BjuA016321	Vps54_N (Vacuolar-sorting protein 54, of GARP complex)
			BjuA016322	Peptidase_S10
			BjuA016324	CKS (Cyclin-dependent kinase regulatory subunit)
			BjuA016325	TPL-binding domain in jasmonate signalling
			BjuA016326	Niemann-Pick C1 N terminus
				Patched family protein
			BjuA016327	Cpn60_TCP1 (TCP-1/cpn60 chaperonin family)
			BjuA016328	ABC2_membrane (ABC-2 type transporter)
				ABC_tran (ABC transporter)
			BjuA016329	UDPGT (UDP-glucuronosyl and UDP-glucosyl transferase)
			BjuA016330	Methyltransf_21 (Methyltransferase FkbM domain)
			BjuA016331	Alpha_L_fucos (Alpha-L-fucosidase)
	B04 (34,761,164)	34,711,164–34,811,164	BjuB000249	SRF-TF (SRF-type transcription factor (DNA-binding and dimerisation domain))
			BjuB000247	CLP1_P (mRNA cleavage and polyadenylation factor CLP1 P-loop)
				Clp1 (Pre-mRNA cleavage complex II protein Clp1)
			BjuB000246	KIP1 (KIP1-like protein)
			BjuB000245	OPT (OPT oligopeptide transporter protein)
			BjuB000243	Transposase_24 (Plant transposase (Ptta/En/Spm family))
	B07 (21,583,568)	21,533,568–21,633,568	BjuB008738	Lectin_legB (Legume lectin domain)
			BjuB008737	UQ_con (Ubiquitin-conjugating enzyme)
			BjuB008735	Choline_transpo (Plasma-membrane choline transporter)
			BjuB008733	Brix (Brix domain)
			BjuB008731	Rad60-SLD_2 (Ubiquitin-2 like Rad60 SUMO-like)
			BjuB008730	NAM (No apical meristem (NAM) protein)

Best hit of BLASTX of gene BjuA003240 to Arabidopsis thaliana genome corresponds to gene AT5G40790. Other names of this gene are AITR3, DIG-LIKE 2 and DIL2. Gene BjuA003241 is involved in nucleic acid binding and RNA–DNA hybrid ribonuclease activity. Alignment of gene BjuA003242 with Arabidopsis genome has 67.28% identity with gene AT4G33790, which encodes for alcohol-forming fatty acyl-CoA reductase (FAR3). According to InterPro database, gene BjuA003242 has three domains viz. IPR005201, IPR013120 and IPR033640. IPR005201 encodes for “glycoside hydrolase, family 85”. IPR013120 encodes for “Fatty acyl-coenzyme A reductase, NAD-binding domain”. IPR033640 encodes for “Fatty acyl-CoA reductase, C-terminal”.

Gene BjuA003244 corresponds to OKI1 gene in A. thaliana (AT4G33760). OKI1 or OKINA KUKI is expressed in the shoot apical meristem (SAM) and encodes a mitochondrial aspartyl tRNA synthetase (AspRS). OKI1 have role in abiotic stress tolerance (Kitagawa et al. 2019). Gene BjuA003245 aligned with Arabidopsis thaliana gene AT4G33740, which encodes for MYB-like protein X. Some MYB transcription factors have been identified to regulate the expression levels of wax-associated FARs under abiotic stresses (Lee and Suh 2015). MYB transcription factors were reported to be involved in the regulation of lipid biosynthesis (Khan et al. 2019). SNP “A06.26228835” on chromosome A06 have five genes in the 50 kb region around SNP. Gene BjuA000608 has 82.62% identity with gene AT3G26130 of Arabidopsis thaliana. AT3G26130 encodes for Cellulase (glycosyl hydrolase family 5) protein. Gene BjuA024420 involved in alpha linolenic acid metabolism. Gene BjuA024421 consists of AP2/ERF domain. Gene BjuA024422 have RNA–DNA hybrid ribonuclease activity. Gene BjuA024423 has protein dimerization activity.

SNP “B04.17559973” have nine genes (BjuB028757 to BjuB028765). BjuB028757 encodes for UNC-50 family protein. UNC-50 family proteins were reported to localize in Golgi bodies in S. cerevisiae and involved in protein trafficking (Chantalat et al. 2003). BjuB028758 encodes for auxin-responsive promoter. BjuB028760 encodes for gibberellin regulated protein. BjuB028762 and BjuB028763 encode for 22-kDa peroxisomal membrane protein. BjuB028764 encodes for transcriptional regulator of RNA polII and BjuB028765 encodes for reverse transcriptase-like protein. Under NS conditions, A04.17533560, B04.34761164 and B07.21583568 have ten, five and six genes respectively in the 50 kb interval around the SNPs. However, no gene was identified in the 50 kb vicinity of SNP A02.22908480. On chromosome A04, BjuA016328 codes for ABC-2 type and ABC transporters and BjuA016329 codes for UDP-glucuronosyl and UDP-glucosyl transferase.

Discussion

Potential donor species for stress tolerance and genetic improvement in Indian Mustard

Wild crucifers harbour the variability for abiotic and biotic stresses (Prakash et al. 2009). B. tournefortii Gouan. (2n = 20), Asian mustard also known as Saharan mustard, has been reported to grown sporadically in a few pockets of arid and semi-arid areas. It has ability to effectively captures available soil moisture and grow fast (Minnich and Sanders 2000). Salisbury (1989) and Prakash and Bhat (2007) have identified B. tournefortii as a potential donor for drought resistance genes. Brassica carinata A. Braun, 2n = 34 (Ethiopian or Abyssinian mustard) possesses excellent drought and heat tolerance, and is a potential crop for dry land cultivation under rainfed conditions (Rakow and Getinet 1997; Jiang et al. 2007). E. cardaminoides (2n = 18, E^cdE^cd) is endemic to rocky places, fields, volcanic rocks and soil, especially in the Micronesian region (Warwick et al. 2000). Its history of evolution makes it a likely source of gene(s) against many abiotic and biotic stresses and capable of gene exchange with the Brassicas (Chandra et al. 2004). Development of B. juncea–E. cardaminoides introgression lines (ILs) with genomic regions associated with Sclerotinia stem rot (Sclerotinia sclerotiorum) resistance has been reported by Rana et al. (2019). These introgression lines, identified for THS tolerance for SOC, are suggested for utilization as donors to broaden the genetic base in breeding for THS in Indian mustard.

Genomic regions associated with variation in seed oil content

SNPs identified under NS conditions could not be associated to those in THS, thereby indicated association of specific genomic regions/QTLs under THS and emphasized the specific need to find marker–trait associations for targeted breeding. Several studies have reported QTLs for SOC on chromosome B04 (Mahmood et al. 2006, Yadava et al. 2012 and Rout et al. 2018) and on chromosome B07 (Yadava et al. 2012; Rout et al. 2018). Cheung et al. (1997) have identified two major QTLs for SOC using RFLP map; one located on linkage group 4 and another on 7. Rout et al (2018) reported one QTL on chromosome A02. SOC-QTL on chromosome A04 under NS condition was first time reported in this study. Moreover, all the previous reported studies have been conducted under normal (non-stress) conditions. However, abiotic stresses caused by extreme temperatures, drought and salinity affect SOC through the activation of various genes in response to these stresses. The activation of genes involved in defence mechanisms against stress may lead to stress tolerance, resulting in higher SOC accumulation compared to plants susceptible to stress. Our study identified genomic regions on chromosomes A01 and B04 that harbour genes involved in abiotic stress tolerance, while chromosomes A01, A04 and A06 contain genes involved in lipid metabolism. SNPs located in these genomic regions show a significant association with variation in SOC. The genes were given in Table 5 and discussed as follows:

(a) Candidate genes involved in abiotic stress tolerance

Large numbers of candidate genes for abiotic stress tolerance were found in the vicinity of SNP “A01.2143072” on Chromosome A01 under THS. Some of these candidate genes were BjuA003238, BjuA003240, BjuA003243 and BjuA003244. BjuA003238 encodes for peroxidase enzyme, which are involved in biosynthesis of phenylpropanoids. Under abiotic stress conditions such as THS, phenylpropanoids biosynthesis is induced in plants resulting in accumulation of various phenolic compounds which, among other roles, have the potential to scavenge harmful reactive oxygen species (Shahivand et al. 2021). Increased synthesis of polyphenols, such as phenolic acids and flavonoids, under abiotic stress conditions, helps the plant to cope with environmental constraints.

Knockout of AITR family genes in Arabidopsis thaliana resulted in the enhanced abiotic stress tolerance without fitness cost (Chen et al. 2021). Such genes are valuable for higher SOC accumulation under THS. Chen et al. (2021) characterized AITRs as suitable candidates for CRISPR/Cas9 editing to improve plant stress tolerance. BjuA003243 encodes proteins involved in phosphatidylinositol signalling system. Gene BjuA003244 corresponds to OKI1 gene in Arabidopsis thaliana (AT4G33760). OKI1 or OKINA KUKI encodes a mitochondrial aspartyl tRNA synthetase (AspRS) which is a key component of the mitochondrial translation apparatus (Kitagawa et al. 2019). Previous studies have reported that mutation in oki1 could affect mitochondrial translation and mitochondrial functions (Robles and Quesada 2017). Role of mitochondria is important in stress response, because, it acts as a powerhouse to produce energy for cells and produces reactive oxygen species (ROS) signals affecting various cellular functions (Huang et al. 2016). BjuA003249 and BjuA003250 which are homologous to AT4G33730 (AtCAPE1) and AT4G33720 (AtCAPE3), respectively are members of CAP (Cysteine-rich secretory proteins, Antigen 5 and Pathogenesis-related 1 protein) superfamily protein encoding small peptides involved in regulation of tolerance to abiotic stress. AtCAPE1 was involved in negative regulation of salt tolerance, while AtCAPE3 was involved in regulation of salt, drought and cold tolerance (Chien et al. 2015; Kim et al. 2021).

Two genes BjuB028762 and BjuB028763 on chromosome B04 under THS were found to be coding for Mpv17/PMP22 proteins. The 22-kDa peroxisomal membrane protein (PMP22) is a hydrophobic integral membrane glycoprotein in all organs of the mature plant (Tugal et al. 1999) with demonstrated roles in cell differentiation and membrane expansion (Suter and Snipes 1995). Mpv17 is a closely related peroxisomal mammalian protein. Its homolog SMY in Saccharomyces cerevisiae has been found to be an integral membrane protein of the inner mitochondrial membrane where it has been proposed to have role in ethanol metabolism and tolerance during heat shock (Trott and Morano 2004).

(b) Candidate genes involved in lipid metabolism

Under THS, this study identified genes (BjuA003242 and BjuA003245) involved in lipid metabolism in the vicinity of genes responsible for abiotic stress tolerance on chromosome A01 (Table 5). BjuA003242 (Fatty acid reductase 3, FAR3, CER4) encodes an alcohol-forming fatty acyl-CoA reductase which have role in lipid biosynthesis (http://brassicadb.cn/#/Annotations/). Fatty acyl reductase (FAR) is a crucial enzyme that catalyzes the NADPH-dependent reduction of fatty acyl-CoA or acyl-ACP substrates to primary fatty alcohols, which in turn acts as intermediate metabolites or metabolic end products to participate in the formation of plant extracellular lipid protective barriers (e.g., cuticular wax, sporopollenin, suberin and taproot wax) (Zhang et al. 2022). They are widely present across plant evolution processes and play conserved roles during lipid synthesis (Zhang et al. 2022). In the vicinity of BjuA003242 (FAR3/ CER4), BjuA003245 is another gene encoding MYB-like protein x (http://brassicadb.cn/#/Annotations/). Lee and Suh (2015) identified MYB transcription factor (AtMYB94) which was induced by salt stress and drought stress, and can activate the expression of AtFAR3/CER4 through direct promoter binding in Arabidopsis thaliana. Khan et al. (2019) reported role of MYB transcription factor (JcMYB1 gene in Jatropha curcas) in enhancing seed oil accumulation and altering fatty acid composition by regulating the expression of endogenous pathway genes in transgenic Arabidopsis and tobacco plants.

Gene BjuA024420 on chromosome A06 involved in alpha-Linolenic acid metabolism (KO id: 10,525; http://brassicadb.cn/#/Annotations/). Alpha-Linolenic acid is important component of seed oil in oilseed crops. Other genes (BjuA024421, BjuA024422, BjuA024423) found in the same region may contribute to regulation of lipid metabolism. Gene BjuA024421 contained AP2/ERF domain, while, gene BjuA024422 and BjuA024423 involved in RNA–DNA hybrid ribonuclease activity and protein dimerization activity, respectively. Xing et al. (2021) studied genome and transcriptome-wide analyses of AP2/ERF and R2R3-MYB transcription factors and proposed potential regulatory network that could control the temperature-induced lipid adjustments in green algae (Auxenochlorella protothecoides). A. protothecoides is known for its potential application in biofuel production. Xing et al. (2021) reported triggering of differential adjustments of lipid pathways with enhanced triacylglycerol accumulation. Under NS, two genes BjuA016328 and BjuA016329 were obtained on chromosome A04. BjuA016328 codes for ABC-2 type and ABC transporters and BjuA016329 codes for UDP-glucuronosyl and UDP-glucosyl transferase. ABC transporters belong to the ATP-Binding Cassette (ABC) superfamily, which uses the hydrolysis of ATP to energize diverse biological systems. Kim et al (2013) have reported role of ABC transporter (AtABCA9) in A. thaliana as a supplier of fatty acid substrates for triacylglycerol (TAG) biosynthesis at the endoplasmic reticulum (ER) during the seed-filling stage.

Conclusion

Overall, our research paper highlights the complexity of the seed oil accumulation process and the impact of high temperature on lipid metabolism in B. juncea. We have identified multiple candidate genes related to abiotic stress tolerance and lipid metabolism that are co-located near significant SNP associations with variability in seed oil content. We have also identified donor genotypes with high SOC under THS conditions for direct utilization in commercial breeding. Further validation of these associations in designated populations will provide a better understanding of their role in B. juncea. Once validated, these marker–trait associations can be used to enhance marker-assisted breeding programs aimed at increasing oil content under high-temperature stress conditions in Indian mustard.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

SKS conceived and designed research. LP and JK conducted experiments. LP and DB analysed data. LP and SKS wrote the manuscript. All authors read and approved the manuscript.

Availability of data and materials

All data generated or analyzed during this study are provided in this published article and its supplementary data files or it will be provided upon a reasonable request.

Declarations

Conflict of interest

The authors declare that no conflict of interest exists.

Informed consent

All authors consent to participate in this manuscript.