Comprehensive evaluation of compound saline-alkali tolerance and gene mining by GWAS in Vicia faba L.

Abstract

Background

Faba bean (Vicia faba L.) is a crucial cool-season legume crop, which is highly valued for its high protein content and key role in crop rotation systems. Considering the increasing threat of soil salinity and alkalinity globally, it is critical to screen germplasm resources with saline-alkali tolerance in faba bean and to identify the underlying genes.

Results

In this study, 12 morphological and physiological traits under compound saline-alkali stress conditions were measured to evaluate saline-alkali tolerance of 240 accessions based on principal component analysis. The results showed that biomass-related traits such as fresh weight of shoot and leaf number had relatively high weights in the evaluation of saline-alkali tolerance at the seedling stage, and 38 highly saline-alkali tolerant accessions were identified. A total of 242 SNPs affecting seedling saline-alkali tolerance were identified in a genome-wide association study of 240 faba bean accessions, with 57 SNPs significantly associated with 7 traits and identified by GLM and MLM models. It was found that 10 genes (such as L-GalLDH, ZAT4, NAC82) overlapped with the reported genes related to saline-alkali tolerance or stress resistance by functional annotation of candidate genes, and their elevated expression in tolerant accessions was validated by RT-qPCR. RT-qPCR indicated tissue-specific differential expression between tolerant and sensitive accessions, with most genes showing higher expression in the tolerant line.

Conclusions

These findings enhance our understanding of the genetic mechanisms governing salt-alkali tolerance in faba bean seedlings and supplies valuable high-tolerance accessions alongside candidate gene resources for breeding programs, thereby laying a solid foundation for subsequent functional validation of these genes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-025-07249-4.

Background

Faba bean, an important cool-season legume, ranks sixth globally in production [1]. It exhibits high yield potential with an average crude protein content of approximately 29%, making it a significant source of plant-based protein, and its amino acid composition meets human dietary requirements [2–4]. Furthermore, faba bean is one of the crops with the highest nitrogen fixation efficiency, reducing the need for nitrogen fertilizers, which provides major benefits for rotation systems and supports sustainable agricultural practices [4, 5]. Its adaptability to a variety of climatic and soil conditions presents obvious advantages over other legume crops. Despite these benefits, the area harvested and production of faba bean still remain lower than those of legumes such as peas and chickpeas, mainly due to unstable yields resulting from biotic and abiotic stresses, highlighting the need for further research to enhance its productivity and stress resistance [6, 7].

Soil salinization and alkalization pose a substantial threat to global food production and severely impede the sustainable utilization of cultivated lands [8, 9]. It is estimated that over 20% of irrigated lands around the world are impacted by excessive salt accumulation. This predicament is exacerbating at an alarming rate, primarily driven by a confluence of factors including unsustainable irrigation practices, inadequate drainage systems, and the far-reaching impacts of climate change [10]. The expansion of saline-alkali soils forces farmers to reduce cultivation areas or even abandon farming altogether, further exacerbating food security issues, particularly in developing countries with limited agricultural resources [8, 10, 11]. 98% of China's saline-alkali land is distributed in the northwest region (Gansu, Shaanxi, and Qinghai), among which the saline-alkali land area in Qinghai Province reaches 3.2 million hectares, accounting for 3.2% of the national total. This area is 5.5 times larger than the cultivated land area of Qinghai Province (586.67 million hectares). Faba bean is an important crop in Qinghai Province, serving multiple purposes for grain, forage, and economic value. It is also a key crop for optimizing the planting structure, increasing export earnings, and improving farmers’ income in Qinghai, representing one of the province’s distinctive agricultural advantages. Therefore, exploring the saline-alkali tolerance of faba bean and breeding elite salt-alkali tolerant varieties are crucial strategies for the efficient utilization and improvement of saline-alkali land [12]. Exploring the saline-alkali tolerance characteristics of crops and breeding superior varieties with saline-alkali resistance are important methods for the utilization and improvement of saline-alkali lands. Faba bean, despite its sensitivity to saline-alkali stress conditions, exhibits a certain degree of tolerance within a specific range of salinity-alkalinity conditions. Furthermore, there exists significant variation in salinity-alkalinity resistance among different genotypes of faba bean [13, 14]. Given the potential of faba bean as a resilient crop, exploring its saline-alkali tolerance characteristics and developing a superior genotype with enhanced resistance could represent a strategic approach for the effective utilization of saline-alkali lands.

A limited number of studies have investigated the harmful effects of saline conditions containing sodium chloride (NaCl), calcium chloride (CaCl₂), and magnesium chloride (MgCl₂) on various genotypes in faba bean, evaluating genotype-specific responses to salinity through emergence rates and early growth characteristics [13–17]. However, these studies utilized a few accessions and didn’t incorporate a comprehensive assessment of trait contributions to saline-alkali tolerance for a robust evaluation of salinity tolerance. While quantitative trait locus (QTL) analysis and genome-wide association studies (GWAS) are commonly employed to identify genes associated with stress tolerance traits such as cold and drought resistance in faba bean [1, 18–20], there is currently no report on QTLs or GWAS related to saline-alkali tolerance.

High concentrations of Na⁺ and Cl⁻ ions, coupled with elevated soil pH, pose significant challenges to plant ion homeostasis and metabolic balance, ultimately leading to impaired growth and reduced productivity[21]. Plants, particularly during the germination stage, exhibit heightened sensitivity to abiotic stress conditions. Even a modest increase in salinity can significantly impair seedling biomass accumulation, highlighting the critical importance of understanding saline-alkali tolerance mechanisms during early development [22]. Comprehensive studies have been conducted to evaluate saline-alkali tolerance and associated genetic determinants in crops such as wheat [23, 24], rice [25, 26], and maize [27, 28]. Given the polygenic and multifactorial nature of saline-alkali tolerance, researchers have employed a suite of phenotypic indices, including relative germination percentage, relative germination index, relative shoot length, relative shoot dry weight per plant, relative root dry weight per plant, relative dry weight per plant, and relative root-to-shoot ratio, to evaluate saline-alkali tolerance and identify key quantitative trait loci (QTLs). Meanwhile, the assessment of crop tolerance based on comprehensive traits requires the application of multivariate statistical methods. These methods include tolerance coefficients, multiple linear regression analysis [29], principal component analysis (PCA) [30], membership function values, comprehensive evaluation (D) values, and cluster analysis, which are widely used to evaluate crop tolerance under stress conditions [31, 32].

In the context of faba bean, while the large genome size (~ 13 Gb) and limited molecular marker resources have historically hindered the application of GWAS, recent advancements in high-throughput sequencing technologies and the release of the faba bean reference genome have made it possible to conduct a comprehensive analysis of the genetic basis of stress response in faba bean [33, 34]. This study aims to evaluate salt-alkali tolerance in faba bean accessions, identify key tolerance genes, and provide genetic resources for breeding resilient varieties, thereby elucidating the molecular mechanisms underlying saline-alkali tolerance in faba bean.

Results

Statistical analysis of 12 phenotype traits under saline-alkali stress

Phenotypic responses to saline-alkali stress exhibited significant differences (P < 0.01) compared to the control condition across all measured traits (Table 1, Fig. 1). Under saline-alkali stress, most growth and biomass traits showed notable reductions in mean values relative to the control. For instance, the average fresh weight of shoot (FWS) decreased from 4.28 g in the control group to 2.30 g under saline-alkali stress, representing a 46% reduction. Dry weight of shoot (DWS) decreased by 60% (from 0.40 g to 0.16 g), fresh weight of root (FWR) decreased by 33% (from 5.70 g to 3.83 g), and leaf area (LA) decreased by 35% (from 8.67 cm² to 5.63 cm²). Additionally, leaf number (LN) and stem diameter (SD) also decreased significantly under saline-alkali stress. A few traits showed little change or even increased under stress. There was no significant difference in dry weight of root(DWR) between the stress and control conditions. The relative chlorophyll content (RCC) of leaves slightly increased under stress (from an average of 38.1 SPAD units to 41.1 SPAD units). The germination rate (GR) under stress conditions only decreased slightly and seeding time (ST) increased slightly compared to the control group.

Table 1.

Statistical analysis of 12 traits in controlled and saline-alkali stress conditions

Traits	Treatment	Mean	SD	CV(%)	Skewness	Min	Max
FWS (g)	Control	4.28	1.30	30.37	0.30	1.48	9.31
	Stress	2.30	0.93	40.43	0.23	0.00	5.52
FWR (g)	Control	5.70	1.83	32.11	1.85	2.59	18.29
	Stress	3.83	1.25	32.64	0.46	0.00	8.16
DWS (g)	Control	0.40	0.14	35.00	0.40	0.08	0.97
	Stress	0.16	0.09	56.25	0.45	0.00	0.46
DWR (g)	Control	0.60	0.17	28.33	2.04	0.29	1.68
	Stress	0.62	0.19	30.65	0.51	0.00	1.53
RCC (SPAD)	Control	38.14	5.08	13.32	0.22	24.47	54.60
	Stress	41.13	7.03	17.09	− 2.29	0.00	59.03
SL (cm)	Control	27.62	4.28	15.50	− 0.23	14.47	40.20
	Stress	20.05	5.39	26.88	− 0.52	0.00	34.87
RL (cm)	Control	24.70	5.47	22.15	0.59	10.20	43.63
	Stress	18.87	5.88	31.16	0.43	0.00	45.70
SD (cm)	Control	0.13	0.02	15.38	0.50	0.07	0.20
	Stress	0.10	0.02	20.00	− 0.81	0.00	0.17
LN	Control	9.65	1.64	16.99	− 0.20	4.00	14.00
	Stress	6.79	1.73	25.48	− 0.81	0.00	11.33
ST (d)	Control	7.00	2.02	28.86	1.02	4.00	15.00
	Stress	7.31	2.36	32.28	0.61	0.00	16.00
GR (%)	Control	0.48	0.32	66.67	0.10	0.00	1.00
	Stress	0.44	0.33	75.00	0.20	0.00	1.00
LA(cm2)	Control	8.67	2.37	27.34	0.69	3.84	18.71
	Stress	5.63	1.83	32.50	− 0.15	0.00	10.96

^aFWS, fresh weight of shoot (g); FWR, fresh weight of root (g); DWS, dry weight of shoot (g); DWR, dry weight of root (g); RCC, relative chlorophyll content (SPAD); SL, shoot length (cm); RL, root length (cm); SD, stem diameter (cm); LN, leaf number; ST, seedling time (days); GR, germination rate (%); LA, leaf area (cm²)

Fig. 1.

Boxplots of the 12 traits under control and saline-alkali stress conditions. Abbreviations wer provided in Table 1

As shown in Talbe 1, the coefficient of variation (CV) of all traits under stress was higher than those in the control group, indicating that various genotypes had significant differences in responses to saline-alkali stress. The CV for FWS reached 40.4%. The CV of the GR and the DWS under stress reached 75.0% and 56.3% respectively. These traits demonstrated substantial genetic variability, which offers a wide range for selecting saline-alkali tolerance.

Comprehensive evaluation (D) of saline-alkali tolerance

Saline-alkali resistance is a complex polygenic trait influenced by multiple genetic and environmental factors, poorly reflected by a single trait. Therefore, we performed correlation and PCA on 12 phenotypic traits to develop a saline-alkali comprehensive evaluation index (D).

Correlation analysis of 12 traits

We constructed a correlation matrix for 12 traits under saline-alkali stress (Fig. 2) and identified strong positive correlations (r > 0.70, P < 0.001) among biomass-related indices, particularly between FWS and FWR. This indicated that genotypes capable of maintaining higher shoot biomass under stress typically exhibited higher root biomass as well. Additionally, SL, LA, and LN displayed highly significant correlations (r ranged from 0.50 to 0.75, P < 0.01). In contrast, RCC exhibited weak correlations with most other traits (|r|< 0.33).

Fig. 2.

Correlation matrix of 12 traits and D-value

Principal component analysis

PCA was used to reduce the dimensionality of the traits. We performed PCA on the standardized values. This analysis grouped the 12 estimated variables into 6 principal comprehensive indices, namely, PC1, PC2, PC3, PC4, PC5, and PC6, which contributed 44.28%, 14.43%, 9.83%, 6.22%, 5.70% and 4.62%, respectively. The cumulative contribution reached 85.06% (Table 2). PC1 was associated with shoot biomass, PC2 with germination-related indicators, PC3 with RCC and FWR, PC4 with root-related indicators, PC5 with RL, and PC6 with SD. The phenotypic interpretation rates of PC1 and PC2 were relatively high, and the cumulative contribution reached 58.68%. Under stress, PC1 and PC2 played a major role in the overall phenotypic variation.

Table 2.

Eigen vectors and percentage of accumulated contribution of principal components

Component	PC1	PC2	PC3	PC4	PC5	PC6
Explained variance (%)	44.28	14.43	9.83	6.22	5.70	4.62
Cumulative variance (%)	44.28	58.68	68.51	74.73	80.43	85.06
FWS	0.92	− 0.12	− 0.10	− 0.07	− 0.15	− 0.03
FWR	0.72	0.26	− 0.50	0.01	0.17	0.02
DWS	0.81	− 0.08	− 0.16	0.06	− 0.22	− 0.27
DWR	0.35	0.66	− 0.42	0.40	0.22	0.01
RCC	0.35	0.43	0.68	0.22	− 0.04	− 0.32
SL	0.86	− 0.05	0.11	− 0.05	− 0.08	− 0.09
RL	0.56	0.23	0.17	− 0.56	0.50	− 0.07
SD	0.71	0.07	0.28	0.19	− 0.07	0.47
LN	0.82	− 0.10	0.12	0.11	0.02	− 0.09
ST	− 0.26	0.81	0.23	− 0.12	− 0.13	0.22
GR	0.50	− 0.54	0.25	0.24	0.36	0.22
LA	0.71	0.05	− 0.08	− 0.31	− 0.37	0.20

Comprehensive evaluation of saline-alkali tolerance and hierarchical cluster

Subordinate function values (μ) of each accession were calculated using Eq. (2). For the same comprehensive index, the saline-alkali tolerance of accessions could be evaluated according to the subordinate function μ value. For PC1, the μ(× 1) values of 8 accessions (T298, T307, T351, T590, T780, T781, T790, and T1060) exceeded 0.9, indicating the highest level of saline-alkali tolerance. Conversely, 3 accessions including T1020, T1099, and T1130 had the lowest μ(× 1) value of 0.00, suggesting the poorest saline-alkali tolerance.

The comprehensive weights for the comprehensive indices were calculated using Eq. (3). The comprehensive weights of the 6 indices (W1, W2, W3, W4, W5, and W6) were 52.04%, 16.95%, 11.56%, 7.31%, 6.71%, and 5.44%, respectively. The comprehensive evaluation D value for saline-alkali tolerance at the seedling stage was calculated using Eq. (4). There were a total of 33 accessions with a D-value greater than 0.7.

Hierarchical clustering based on the D values classified the examined 240 accessions into five groups (Fig. 3). Cluster Ⅴ, the highly saline-alkali-tolerant group, included 38 accessions, such as T255. These accessions exhibited significantly higher biomass and chlorophyll content compared to the other groups. Cluster Ⅳ, the saline-alkali-tolerant group, included 76 accessions such as T123. Cluster Ⅲ, the moderately saline-alkali-tolerant group, included 59 accessions such as T119. Cluster Ⅱ, the low saline-alkali-tolerant group, included 55 accessions such as T109. Cluster Ⅰ, the saline-alkali-sensitive group, included 12 accessions such as T163.

Fig. 3.

Cluster analysis for saline-alkali tolerance in 240 faba bean accessions based on D values

We analyzed the correlation between 12 phenotypic indices and the D value. With the exception of ST, the other 11 indices exhibited extremely significant positive correlations with the D value. Notably, the correlation coefficient between SD and D reached 0.8, indicating a strong positive relationship.

Population genetic structure

The population structure of 240 faba bean accessions was analyzed using the Admixture software, and the CV error was employed to determine the optimal number of subgroups. The results indicated that the CV error reached its minimum value at K = 4(Fig. 4b), suggesting that the accessions could be partitioned into four distinct subgroups(Fig. 4a): Subpop I, Subpop II, Subpop III, and Subpop IV. Accessions with membership probabilities ≥ 0.50 were considered to belong to the same group. Among the 240 accessions, 51 were assigned to Subpop I, 81 to Subpop II, 42 to Subpop III, and 31 to Subpop IV, while the remaining accessions were classified as a mixed group. Notably, accessions from diverse geographical origins or ecological types were grouped within the same subgroups, which implies the occurrence of gene flow or introgression among the faba bean accessions.

Fig. 4.

Population genetic structure analysis based on PCA and admixture results. a PCA of 240 accessions. b Relationship between K and CV error. c The structure of the 240 accessions based on Structure software when K = 4

To better elucidate the genetic relationships among the accessions, PCA was performed using PLINK software on the 240 accessions, and the results were visualized as a three-dimensional scatter plot (Fig. 4c). The first three principal components, PC1 (35.22%), PC2 (14.52%), and PC3 (10.96%), collectively captured 60.70% of the total genetic variation. The results revealed that the 240 accessions were clearly divided into four distinct clusters, with some overlap observed among certain accessions. These findings aligned with the results of the population structure analysis, thereby demonstrating the reliability of the groups classification P < 1.0 × 10⁻⁴.

GWAS for saline-alkali tolerance indexes

To elucidate the genetic basis of seedling-stage saline–alkali tolerance in faba bean, we integrated phenotypic and genotypic data and performed GWAS using both the general linear model (GLM) and the mixed linear model (MLM). We used a stringent genome-wide cutoff of − log10(P) > 4.0 (i.e., P < 1.0 × 10⁻⁴). With population structure (Q) and kinship (K) included as covariates, the union of GLM and MLM results yielded 242 genome-wide significant single-nucleotide polymorphisms (SNPs), of which 57 were detected by both models. Eighteen of these SNPs were simultaneously associated with four or more traits, indicating that they may play pivotal roles in the faba bean response to saline–alkali stress and are promising candidates for further functional investigation.

Among all significant loci, the greatest number (145) were associated with GR, whereas fewer than 10 loci were linked to FWR and RL. Manhattan and Q–Q plots for these associations are provided in Supplementary Figures S1 and Figures S2.

Candidate gene analysis

Based on the Hedin/2 reference genome in faba bean, a total of 103 genes were predicted, which are distributed across multiple chromosomes. Among them, 10 candidate genes closely related to saline-alkali tolerance traits were identified. The gene Vfaba.Hedin2.R1.3g023640, located on Chromosome 3, was annotated as L-GalLDH. This gene encodes L-galactono-1,4-lactone dehydrogenase, an enzyme that is of critical importance in the synthesis pathway related to maintaining redox balance [35, 36]. Vfaba.Hedin2.R1.2g026040 was annotated as ZAT4, a member of the C2H2 zinc finger TF family. Previous studies have reported that ZAT4 can enhance plant tolerance to osmotic stress and salt stress via the abscisic acid (ABA) signaling pathway [37, 38]. Vfaba.Hedin2.R1.2g148560 was annotated as NAC domain-containing protein 82. NAC transcription factors are one of the largest families of plant transcription factors and have been widely proven to be involved in abiotic stress adaptation [39–41].

The remaining candidate genes were predominantly involved in ion transport, signal protein synthesis, and the synthesis of protective enzymes associated with the antioxidant pathway. These functions are all integral to the plant’s ability to cope with saline-alkali stress, as they help maintain cellular ion balance, transmit stress signals, and detoxify ROS. In summary, these 10 candidate genes (Table 3) provide valuable insights into the genetic basis of saline-alkali tolerance in faba bean and serve as potential targets for further genetic improvement and functional studies.

Table 3.

The information of significant association loci related to 12 candidate genes

ID	Trait	Marker	Model type	Chr	p	Gene name	Gene description
1	ST	dou_TRINITY_DN47983_c0_g1_503	GLM	chr3	3.92E-05	Vfaba.Hedin2.R1.3g023640	L-galactono-1,4-lactone dehydrogenase, mitochondrial
2	ST	dou_TRINITY_DN50263_c0_g1_698	GLM	chr2	2.35E-05	Vfaba.Hedin2.R1.2g026040	Zinc finger protein ZAT4
			MLM		8.27E-05
3	FWS	hua_TRINITY_DN149364_c1_g1_436	GLM	chr1	5.15E-05	Vfaba.Hedin2.R1.1g188000	1-aminocyclopropane-1-carboxylate oxidase homolog 1
4	GR	ye_TRINITY_DN135934_c6_g2_1000	GLM	chr3	4.85E-05	Vfaba.Hedin2.R1.3g108200	NAC domain containing protein 50
5	GR	ye_TRINITY_DN125437_c2_g1_365	GLM	chr3	1.63E-05	Vfaba.Hedin2.R1.3g108200	NAC domain containing protein 50
6	GR	ye_TRINITY_DN128763_c0_g1_87	GLM	chr4	6.23E-05	Vfaba.Hedin2.R1.4g007000	Sucrose transport protein SUC3
			MLM		8.91E-05
7	GR	ye_TRINITY_DN142350_c3_g1_1444	GLM	chr4	6.74E-05	Vfaba.Hedin2.R1.4g007000	Sucrose transport protein SUC3
			MLM		9.00E-05
8	GR	ye_TRINITY_DN129792_c1_g2_262	GLM	chr3	1.55E-05	Vfaba.Hedin2.R1.3g084000	Gibberellin 20 oxidase 2
			MLM		7.82E-05
9	GR	ye_TRINITY_DN129792_c1_g2_293	GLM	chr3	9.00E-06	Vfaba.Hedin2.R1.3g084000	Gibberellin 20 oxidase 2
			MLM		7.94E-05
10	GR	ye_TRINITY_DN130469_c4_g4_600	GLM	chr3	1.17E-07	Vfaba.Hedin2.R1.3g214960	K(+) efflux antiporter 6
			MLM		3.39E-06
11	GR	ye_TRINITY_DN136702_c1_g3_568	GLM	chr2	2.96E-05	Vfaba.Hedin2.R1.2g082800	Mitogen-activated protein kinase kinase kinase 1
12	BL	ye_TRINITY_DN146501_c1_g2_1391	GLM	chr3	1.37E-05	Vfaba.Hedin2.R1.3g036920	E3 ubiquitin-protein ligase PUB22
			MLM		3.82E-05
	DWR		GLM		6.00E-05
	LN		GLM		7.88E-05
	RCC		GLM		1.69E-09
			MLM		1.70E-07
	SD		GLM		7.34E-06
			MLM		2.76E-05
	ST		GLM		3.03E-05
13	GR	ye_TRINITY_DN150763_c2_g1_517	GLM	chr1	1.99E-05	Vfaba.Hedin2.R1.1g093640	Chloride channel protein CLC-c
			MLM		7.22E-05

Expression analysis of candidate genes by qRT-PCR

The RT-qPCR results indicated significant differences in expression among the candidate genes across various tissues (Fig. 5). The Vfaba.Hedin2.R1.1g093640 exhibited significantly higher expression levels in the roots and leaves of the tolerant accession T351 compared to the sensitive accession T1099 (P < 0.001), whereas its expression was significantly lower in stems (P < 0.05). In contrast, the Vfaba.Hedin2.R1.2g026040 showed significantly higher expression in the stems of sensitive accession T1099 (P < 0.0001), while roots and leaves displayed higher expression levels in the tolerant accession T351 (P < 0.01).

Fig. 5.

Tissue-specific expression of candidate genes (T351 vs. T1099)

Other candidate genes such as Vfaba.Hedin2.R1.3g023640, Vfaba.Hedin2.R1.3g214960, Vfaba.Hedin2.R1.2g082800, Vfaba.Hedin2.R1.1g188000, Vfaba.Hedin2.R1.4g007000, and Vfaba.Hedin2.R1.3g084000 generally exhibited higher expression levels in the tolerant accession T351 across most tissues (P < 0.05 or P < 0.01), suggesting their significant roles in saline-alkali tolerance mechanisms during the seedling stage in faba bean.

It is noteworthy that some genes, such as Vfaba.Hedin2.R1.3g108200, did not show significant expression differences between the two accessions in young leaves, though significant differences were observed in other tissues. This highlights the importance of tissue-specific gene expression patterns in studying saline-alkali tolerance mechanisms.

Overall, the RT-qPCR results validated the gene expression differences identified by GWAS analysis and further supported the importance of these candidate genes in the saline-alkali stress response of faba bean seedlings. These findings provide valuable references for deeper understanding of the molecular mechanisms underlying saline-alkali tolerance and facilitate molecular breeding in faba bean.

Discussion

In China, saline-alkali land accounts for 25% of agricultural cultivated areas, particularly in the northwest, north, and parts of the northeast regions [42]. These lands remain underutilized due to soil constraints. With continuous changes in land use patterns, the area of saline-alkali land has been increasing annually [43]. Under such circumstances, investigating crop responses to saline-alkali stress becomes critically important. As a vital cool-season legume crop [1], the saline-alkali tolerance of faba bean directly determines its cultivation feasibility in saline-alkali regions. The seedling stage represents the most vulnerable phase in the crop lifecycle, during which environmental stresses exert the most pronounced impacts [44–46]. Therefore, analyzing phenotypic changes at this stage under abiotic stress can effectively reflect crop stress tolerance. Previous studies have demonstrated that biomass-related traits of faba bean seedlings, including FWS, SL, and LA, are significantly reduced under saline-alkali stress [47, 48]. Investigating the phenotypic responses of faba bean seedlings to saline-alkali stress will contribute to elucidating the underlying tolerance mechanisms and provide theoretical support for breeding saline-alkali tolerant faba bean cultivars.

Response of faba bean seedling phenotypes to saline-alkali stress

In this study, saline-alkali stress was applied to faba bean at the seedling stage, resulting in significant phenotypic changes among the tested accessions. Comparative analysis between the control and stress groups further revealed the detrimental effects of saline-alkali stress on seedling growth. Saline-alkali stress inhibited water uptake and developmental processes in plants, as evidenced by reductions in biomass-related traits, including FWS, FWR, and seedling length. These findings are consistent with the results reported by Ziche et al. [49] and Diab et al. [50]. In addition, saline-alkali stress exerted inhibitory effects on seed germination, manifested by reduced germination rates and prolonged germination periods, which also aligns with the observations of Diab et al. [50]. In agreement with our findings, other leguminous crops have also demonstrated suppressed germination and inhibited seedling growth under saline-alkali stress. For example, in the leguminous forage crop alfalfa, high concentrations of alkaline salts significantly inhibit seed germination and early seedling growth [70]. Similarly, chickpea seedlings exhibit significant reductions in shoot and root growth, as well as decreased biomass accumulation, under alkali stress conditions [71]. Such inhibitory effects on seed germination and seedling development are also observed in cereal crops. For instance, rice is highly sensitive to saline-alkali conditions, with substantial reductions in germination rate and severe inhibition of seedling growth [72]. Taken together, saline-alkali stress severely impairs both biomass (fresh weight and seedling length) and germination characteristics (germination rate) in plants. Therefore, shoot biomass and germination-related indices should be considered as key indicators for evaluating the saline-alkali tolerance of faba bean varieties.

Comprehensive evaluation for saline-alkali tolerance of faba bean accession at the seedling stage

Saline-alkali tolerance is a complex, polygenic trait in crops, and any single phenotypic index is inadequate for a comprehensive assessment [51–53]. Here, we implemented a four-tier integrative framework—“SATC → PCA → membership function → D-value clustering”—to systematically evaluate the 240 accessions. Principal-component analysis showed that the first two components (PC1 + PC2) explained 58.68% of the total variation, driven chiefly by biomass-related traits (FWS, FWR, SL) and the germination trait (emergence time), underscoring their pivotal role in saline-alkali tolerance. Tavakoli et al. [54] and Benabderrahim et al. [55] used a comprehensive analysis of multi-index results to better reflect the salt tolerance of alfalfa seedlings. By comparison, Yu et al. [32] obtained 64.8% explanation using seven indices across 15 alfalfa genotypes; in our study, the first two components alone accounted for 58.7%, and the first six components cumulatively explained 85% of the variance. Based on the comprehensive D-value, the 240 accessions were partitioned into five clusters; Cluster Ⅴ displayed the highest saline-alkali tolerance and comprised 38 accessions (e.g., T255, T289), furnishing a richer core gene pool for subsequent molecular breeding. It should be emphasized, however, that evaluation at the seedling stage alone cannot fully represent tolerance across the entire life cycle; therefore, more exhaustive assessments spanning all growth stages are warranted for rigorous validation.

Identification of saline-alkali tolerance-related loci and candidate genes in faba bean

In this study, a GWAS identified 242 significant SNP loci associated with saline-alkali tolerance in faba bean, 57 of which were consistently detected by both GLM and MLM. Functional annotation further prioritized 10 candidate genes strongly associated with saline-alkali tolerance traits. Among these, the L-GalLDH gene (Vfaba.Hedin2.R1.3g023640) plays a critical role in the antioxidant response pathway, encoding an enzyme pivotal in ascorbic acid (AsA) biosynthesis by converting L-galactono-1,4-lactone to AsA. AsA, a major non-enzymatic antioxidant in plants, mitigates oxidative stress by scavenging reactive oxygen species (ROS) under stress conditions, thus profoundly influencing salt tolerance [56, 57]. Studies have demonstrated salt stress significantly affects AsA content; for instance, Zhang et al. [58] reported enhanced salt tolerance in rice overexpressing L-GalLDH, correlated with increased ascorbate content. Similarly, Alharby et al. [57] found mild salt stress slightly induced ascorbate levels in soybean leaves, whereas severe salt stress caused decreased ascorbate content and disrupted antioxidant systems. Although direct functional validation of L-GalLDH in legumes is currently limited, insights from Arabidopsis research by Smirnoff et al. [59] showed that RNAi-mediated suppression of L-GalLDH led to reduced ascorbate levels and impaired plant growth, underscoring its crucial role in plant oxidative stress tolerance. The ZAT4 gene (Vfaba.Hedin2.R1.2g026040) belongs to the C2H2 zinc finger transcription factor family, known for involvement in responses to drought, salt, and cold stresses in plants [60]. The soybean gene GmZAT4, homologous to AtZAT4 from Arabidopsis, is significantly induced under stress conditions and predominantly expressed in roots, suggesting its essential role in root-based stress signaling [38]. Sun et al. [61] reported that Arabidopsis plants overexpressing GmZAT4 exhibited significantly enhanced salt tolerance, accompanied by increased activities of antioxidant enzymes such as ascorbate peroxidase (APX) and superoxide dismutase (SOD). Thus, ZAT4 and related zinc finger transcription factors represent valuable candidate genes for improving crop salt tolerance through molecular breeding. Additionally, the NAC domain-containing protein 82 gene (Vfaba.Hedin2.R1.2g148560), a member of the extensively studied NAC transcription factor family implicated in abiotic stress responses, enhances saline-alkali tolerance by maintaining cellular osmotic homeostasis and augmenting antioxidant capacity. Bokolia et al. [62] conducted a genome-wide identification of NAC transcription factors in oat, revealing significant expression profiles under salt stress conditions, providing a reference for further NAC family research in legumes. Other candidate genes also likely play integral roles in saline-alkali tolerance mechanisms. Subsequent RT-qPCR analyses (Fig. 5) confirmed markedly higher expression levels of these candidate genes in the tolerant accession T351, further validating their essential functions under saline-alkali stress. These findings deepen our understanding of saline-alkali tolerance mechanisms in faba bean and offer valuable gene resources for future molecular breeding and functional genomics research.

Conclusions

This study evaluated the saline-alkali tolerance of 240 faba bean accessions at the seedling stage. The comprehensive assessment identified 38 accessions (e.g., T255, T289) exhibiting strong tolerance under saline-alkali stress, which represent critical genetic resources for breeding saline-alkali tolerant cultivars. Genome-wide association study revealed 242 significant SNP loci associated with saline-alkali tolerance. Functional annotation further prioritized 10 candidate genes (e.g., L-GalLDH, ZAT4, NAC82) involved in faba bean saline-alkali tolerance. whose elevated expression in tolerant accessions was subsequently confirmed by RT-qPCR. These genes are functionally enriched in antioxidant responses, osmotic regulation, and signal transduction pathways, providing actionable targets for molecular design breeding.

In summary, this study advances the understanding of the genetic mechanisms governing saline-alkali tolerance in faba bean and provides valuable gene resources for breeding programs. Future efforts focusing on functional validation of these candidate genes and the application of gene editing technologies are expected to enhance faba bean performance in saline-alkali regions, ultimately facilitating its widespread cultivation in marginal lands.

Accessions and methods

Plant accessions

A panel of 240 faba bean accessions from different countries was phenotyped for seedling-stage saline-alkali tolerance. Of these, 183 were from China, and 57 were from other countries. These 240 accessions of faba been accessions resources used in this research come from Qinghai Academy of Agricultural and Forestry Sciences, and all accessions resources were cultivated at the Academy of Agriculture and Forestry Sciences Qinghai University.

The panel had been genotyped with the faba bean 130 K targeted next-generation sequencing SNP genotyping platform[63]. From this SNP dataset, 30,338 SNPs were selected as high-quality SNPs with minor-allele frequency (MAF) greater than 0.05 and maximum missing rate less than 20%.

Experimental procedures and phenotypic characterization

Each treatment was arranged with three biological replicates, and five seeds were used for each accession to assess seedling (the two-true-leaf stage) saline-alkali tolerance. The specific developmental stage was determined according to the standards proposed by Meier U et al. [64]. Seeds were sown in pots and grown in an artificial climate chamber with an 18 h/6 h light/dark photoperiod, day/night temperatures of 22 °C/17 °C, and approximately 50% relative humidity. A saline-alkaline solution composed of NaCl, sodium carbonate (Na₂CO₃), and sodium bicarbonate (NaHCO₃) with a 9:1:1 ratio was used to irrigate the stress group. The final concentration of 100 mmol/L. Throughout the experiment, the pH of the saline-alkali solution was consistently maintained at 9.2. The control group was irrigated with deionized water. A randomized experimental design was employed during cultivation to minimize experimental error.

A total of 12 morphological and physiological traits with the two-true-leaf stage. These traits include FWS, FWR, DWS, DWR, RCC in leaves, RL, SL, SD, LN, GR, ST, and LA. Standard protocols were used for all measurements, with three biological replicates for each measurement.

Statistical analysis

Refer to the comprehensive evaluation (D) value of crop stress tolerance evaluation to conduct statistical analysis on the data [32, 65, 66]. The saline-alkaline tolerance coefficient (SATC) of each of the above indices was calculated.

1

SATC_ij represents the saline-alkaline tolerance coefficient of index (j) for accessions (i); X_ij(control) and X_ij(stress) denote the values of the index for the accessions evaluated under ddH₂O and saline-alkaline treatments, respectively.

The membership function value μ was calculated using Eq. (2).

2

The X_i is the ith comprehensive index; X_imax and X_imin represent the maximum and minimum values for the ith comprehensive index of each material, respectively.

In Eq. (3) the weight function W_i was calculated and represents the relative importance of the ith comprehensive index for a accession.

3

P_i represents the contribution to the ith comprehensive index.

In Eq. (4) the comprehensive evaluation parameter for saline-alkaline tolerance resilience (D) for each cultivar was calculated to identify the saline-alkaline tolerance capacity of the different accessions.

Then, hierarchical cluster analysis was also used to evaluate salt tolerance.

4

μ(X_i) is the membership function value of ith comprehensive index; W_i represents the relative importance of the ith comprehensive index.

SPSS26.0 and R4.3.2 software were used to conduct correlation analysis, PCA, membership function analysis, and cluster analysis. R4.1.3, Origin 2022, and GraphPad Prism 9 were used to draw the images required in this study.

Population structure analysis

To estimate the population structure and phylogenetic relationships within the panel of faba bean, the Admixture software was run from 2 to 10 to analyze 30,338 markers. A tenfold CV scheme was repeated 10 times for each value of K. The optimal grouping was determined by identifying the K-value associated with the minimum CV error, which was visualized through a curve plotting the CV error against the K-value. Accessions with membership probabilities ≥ 0.50 were considered to belong to the same group. We then utilized R4.1.3 to visually represent the admixture proportions. For PCA, the PLINK v.1.9 software was applied, setting the number of principal components (PCs) equal to the number of samples[67]. The resulting eigenvector plot was generated using the ggplot2 package in R, following the methodology described by Wickham[68].

Genome-wide association analysis

GWAS of traits related to saline-alkali tolerance were conducted with both the GLM and MLM, integrating the Q matrix generated by Admixture and the K matrix produced by TASSEL 5. The significance of associated markers was evaluated using P-values, with a stringent threshold of -log10(P) > 4.0 applied to identify significant associations. This threshold was determined based on Bonferroni correction to account for multiple testing. Finally, the CMplot package in R software was used to visualize the results of the GWAS through a Manhattan plot and a quantile–quantile (Q–Q) plot. Functional annotations of candidate genes within the identified regions were retrieved using the Hedin/2 reference genome (https://projects.au.dk/fabagenome/genomics-data).

Candidate gene analysis and RT-qPCR

To validate the expression levels of candidate genes identified through GWAS analysis, RT-qPCR experiments were conducted. Based on comprehensive evaluation results, two representative accessions were selected: one tolerant (T351) and one sensitive (T1099). Seedlings of these two accessions were sampled from different tissues (roots, stems, and leaves) at 14 days under both saline-alkali stress and control conditions, with three biological replicates for each sample. Total RNA was extracted using a plant RNA extraction kit (Tiangen, Beijing, China) and was reverse-transcribed using a PrimeScript™ RT reagent Kit with gDNA Eraser (Tiangen, Beijing, China). RNA quality was verified by 2% agarose gel electrophoresis using the UVITEC Cambridge FireReader V10 gel imaging system. Gene-specific primers were designed using Primer Premier 5.0 software based on the sequences from the Hedin/2 reference genome. RT-qPCR assays were performed on a Roche LightCycler480 II (Roche Molecular Biochemicals, Mannheim, Germany) according to the protocol described by Yao et al. [69]. The ACTIN gene from Vicia faba was used as the internal reference to normalize gene expression levels. Each reaction was conducted with three technical replicates. Relative gene expression levels were calculated using the 2 − ∆∆Ct method, and statistical analysis was performed using GraphPad Prism 9 software.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

FWS

Fresh weight of shoot

FWR

Fresh weight of root

DWS

Dry weight of shoot

DWR

Dry weight of root

RCC

Relative chlorophyll content

RL

Root length

SL

Shoot length

SD

Stem diameter

LN

Leaf number

GR

Germination rate

ST

Seeding time

LA

Leaf area

GWAS

Genome-wide association studies

SNP

Single nucleotide polymorphism

GLM

General linear model

MLM

Mixed linear model

AsA

Ascorbic acid

QTL

Quantitative trait locus

PCA

Principal component analysis

CV

Coefficient of variation

Author contributions

X.P. contributed to software, data analysis, writing, and editing. Z.S. contributed to software, data analysis, and editing. X.Z. contributed to software, supervision, data analysis, and writing. X.W. contributed to the investigation, review, and editing. D.Z. contributed to software, investigation, and writing. C.T. contributed to software, data analysis, and writing. H.Z. contributed to experimental design, methodology, software, investigation, data analysis, and writing the original draft. W.H contributed to software, data analysis. L.Y. contributed to supervision, funding acquisition, writing, review, and editing. The author(s) read and approved the final manuscript.

Funding

The Kunlun Talent-Advanced Innovation and Innovative and Entrepreneurial Talent Project of Qinghai Province (2024003); The Key Research and Development and Transformation Plan of Qinghai Province (2022-NK-109); National Natural Science Foundation of China (NSFC,42267008) and the China Agriculture Research System of MOF and MARA (CARS-08).

Data availability

All the accessions were provided by the Legume Research Group of the Qinghai Academy of Agricultural and Forestry Sciences. All the data generated were listed in the supplementary tables and figures. Data is provided within the manuscript or supplementary information files.

Declarations

Ethics approval and consent to participate

All procedures were conducted following the guidelines.

Consent for publication

Not applicable.

Competing interest

The authors declare no competing interests.