Introduction

Legumes chosen for seeds as a portion of food are known as pulses. These are cool seasons, annual species. Pulses are an alternate dominating dietary source of protein other than meat [1]. Among legumes, chickpea (Cicer arietinum L.) occupies the central stage for world production, imports, and exports, because of its extraordinary nutritional values. Globally, chickpea is the third most important food legume after dry bean (Phaseolus vulgaris L.) and field pea (Pisum sativum L.) [2]. It belongs to the family Fabaceae [3] and the tribe Cicereae.

Chickpea seeds are rich in protein, have low fat & sodium, and zero cholesterol. They are also a great source of soluble and insoluble fibers, carbohydrates, vitamins, folate, and minerals [4, 5]. Because of its highly balanced composition of amino acid and protein bioavailability, it is also considered a worthy source of dietary protein [6]. In addition to its cheap availability because of its low price, its consumption is also related to the avoidance of cardiovascular disorders, handling type-2 diabetes, and reducing the levels of LDL cholesterol. The dietary insoluble fibers are related to the prevention of colon cancer as well as soluble fibers help in the maintenance of healthy body weight [4]. In addition, vitamins [tocopherol (both γ and α)], folic acid, riboflavin (B2), pantothenic acid (B5), pyridoxine (B6), and carotenoids such as β-carotene, lutein, cryptoxanthin, and zeaxanthin are also in ample amount in chickpea [7].

The origin and cultivation of chickpea started in the Middle East era ~ 7 millennia ago [1]. Since then chickpea has found its place in fields of temperate and semi-arid states of Asia, Europe, Australia, and North America. Currently, chickpea is grown in nearly 52 countries [8]. Despite that, approximately three-quarters of the global chickpea producers are coming from South Asian countries [9] where India is by far the largest producer providing 75% of the world’s chickpea yield [10], followed by Turkey, Pakistan, and Iran with approximately 7%, 6%, and 4% world supply, respectively.

Two main types of cultivated chickpea are “Desi” and “Kabuli” where “Kabuli” genotypes are characterized by a larger, rounder, and cream-colored seeds (mainly grown in North Africa, West Asia, North America, and Europe) and “desi” genotypes have smaller, angular-shaped and dark colored seeds with wrinkled seed coat, mostly grown in Asia and Africa [11]. The flower color of the “desi” type is pink with a semi-erect or semi-spreading growth type and is grown mainly in Ethiopia and the Indian subcontinent. The flower color of the “Kabuli” type is a white semi spreading growth habit and cultivated in temperate regions of the world [4].

The genus Cicer contains (nine annual and 34 perennial) species. According to morphological characteristics, life span, and geographical allocation, Cicer species are organized into four Sect. [12]. Section Cicer (= Monocicer M.G. Popov) is characterized by annual, small-flowered species with firm, erect to inclined or prostrate stems and imparipinnate leaves, or the rachis ending in a tendril (carries eight annual species); Section Chamaecicer M.G. Popov contains shrubby species with thin, creeping branches and 3–7 leaflets per leaf (carries two perennial species viz. C. chorassanicum, C. incisum); Section Polycicer M.G. Popov contains the perennial, rather large-flowered species with imparipinnate leaves or the rachis ending in a tendril (contains 23 perennial species); and Section Acanthocicer M. G. Popov includes perennial species with a large flower, spiny leaf rachis and spiny calyx teeth (carries seven woody perennial species). Among these four sections, “Cicer” (containing C. reticulatum, C. echinospermum, C. pinnatifidum, C. arietinum, C. bijugum, C. judaicum, C. yamashitae, and C. cuneatum) has attracted considerable attention of breeders [12].

Agricultural productivity is faced with dual challenges in the form of an increasing human population and irregular climate changes [13]. Ecosystems and environments have been altered because of changes in rainfall patterns and seasons. This situation dictates that joint efforts may be made to enhance the productivity of breeding programs toward accelerated climate smart varietal development [10, 14]. The situation becomes even more challenging in the case of chickpea because of its narrow genetic diversity that impedes the improvement and characterization of molecular markers for any trait [15]. To add to it, genetic bottlenecks are created [16] because of the use of few successful varieties for future breeding. Hence, to hasten the development of new varieties with increased annual yield, selection and use of diverse genotypes in a breeding program are required to create genetic variability [17].

The adaptive behavior of organisms is determined by the genetic diversity that it contains which is developed over a vast period of evolution [18]. The raw material in the form of diversity is the basis of introducing desired alleles in modern cultivars. This diversity can be expanded by using wild and cultivated relatives of a given species. The first step towards diversity enhancement is its assessment. Diversity can be gauged by taking into account the pedigree data [19] passport data, morphological data [20], biochemical data [21], storage proteins [22], and DNA-based marker data [23]. Among them, the use of storage protein profiling has been widely adopted to understand the latent evolutionary aspects of different crops [24].

One of the most economical and widely adopted biochemical methods to analyze the genetic relationships in germplasm is Sodium Dodecyl Sulphate Polyacrylamide Gel Electrophoresis (SDS-PAGE) [25]. Seed storage proteins have been identified as potent markers for; (a) discerning the diversity between and within species; (b) domestication for genetic resources and conservation and breeding, (c) to ascertaining the genetic relationships, and (d) the improvement of important agronomic and economic traits in crop plants [24]. The electrophoretic separation of seed storage protein has also been used to investigate the variations in the genome and to identify plant varieties [26]. Seed proteins are not susceptible to ecological fluctuation; the banding pattern of seed protein is highly stable which encouraged the identification of cultivars. Protein banding patterns also serve as an additional tool for cultivar identification or when the cultivars are to be patented [27].

Owing to the importance of chickpea and the selection and use of diverse genotypes in a breeding program with the larger aim to create genetic variability, we initiated this study to dissect the genomic diversity of 436 accessions of nine annual Cicer species by seed storage profiling using SDS-PAGE technique. The general objectives were to find out the polymorphism in seed storage proteins of the germplasm and to identify markers that could successfully differentiate various genotypes.

Materials and methods

Plant materials

This research was conducted in Plant Breeding and Genetics Division, Nuclear Institute for Agriculture and Biology (NIAB) Faisalabad. This study was accomplished on a set of 436 accessions of 9 annual Cicer species. The germplasm was acquired from Australian Grain Genebank is originated from 47 different countries. It comprises 305 cultivated and 131 wild type chickpea genotypes. Passport information of germplasm used in this study is provided in Table S1. The Pakistani accessions were available at NIAB.

Methods

Sample preparation

One seed from each accession was ground to a fine powder in pastel & mortar. The measured quantity of powder from each ground sample (0.05 g) was solubilized in 1ml Tris Urea Extraction Buffer [0.5 M Tri buffer (pH8), Urea (30.3 g/dry), SDS (10%), and 1ml Beta-merceptoethanol in total volume of 100ml] employing modified standard protocol for SDS-PAGE [28] in eppendorf tubes and centrifuged for 10 min at 14,000 rpm in a micro-centrifuge machine. The supernatant was collected and stored at −20 °C for subsequent analysis.

Protein estimation and SDS-PAGE

A dye-binding assay by [29] was carried out to estimate the protein concentrations in the supernatant. A volume of 2 µl of supernatant and 8 µl of sample preparation buffer i.e. extraction buffer mixed with tracking dye Bromophenol Blue (BPB) were loaded into wells and 5 µl of pre-stained broad range protein marker (Bio-Rad Catt log#161–0317) was loaded in a separate well. Electrophoresis was performed using a Mini-PROTEAN BIO-RAD electrophoresis system at a constant voltage (35 V) for 3–4 h in 1X running buffer (For 5X: 15 g of Tris-base, 72 g Glycine, and 5 g SDS in 1 L of distilled water).

After completion of SDS-PAGE, the gel was stained for 5–6 h on the shaker in a staining solution containing H2O, ethanol, and acetic acid with a ratio of 5:4:1, respectively, and 1 g of Coomassie brilliant blue R250 dye. Later on, the gel was destained in a destaining solution overnight. Destaining solution was the same as the staining solution without Comassie Brilliant Blue R250 dye. The gel picture was captured and stored for further data analysis.

Band scoring and data analyses

Each lane exhibited protein bands from each sample representing one genotype. Prestained marker bands were set as a landmark for the estimation of the molecular weight of protein bands on the gel. All the sample bands were compared to landmarks using the software PyElph 1.4. The analysis of images through software allowed easy and rapid determination of numbers and molecular masses of protein bands resolved on an SDS-PAGE gel. Furthermore, scoring was performed and electropherograms were produced by recording the presence (1) or absence (0) of each band.

Cluster analyses

Cluster analysis was performed in R Studio (version 2021.09.0 Build 351) using the packages “cluster” and “factoextra” which are equipped with different methods of clustering. K-means clustering was performed using the function “fviz_cluster (k2, data = X)” where k2 is defined by the user and symbolizes the presence of two sub-groups in the data “X”. The elbow method was implemented using the function, “fviz_nbclust(X, kmeans, method= “wss”)” where “wss” was defined as within the cluster sum of squares. The silhouette method was implemented using the function “fviz_nbclust(X, kmeans, method= “silhouette”)” where “silhouette” determines the quality ,of a clustering i-e, how well a value fits within its group and a high mean silhouette width is proportional to a good clustering. The “Gap statistic” method was computed using the “clusGap” function which results in providing a gap statistic and the standard error of the output. The following function was used: “gap_stat<-clusGap(X, FUN = kmeans, nstart = 25, K.max = 50, B = 50)” where “K.max” is defined as the maximum number of sub-groups in the dataset. The results were visualized using the function “fviz_cluster(gap_stat)”. After defining the optimal number of clusters (which in our case was 14), k-means clustering was performed by defining K = 14 using the function “optimal<-kmeans(X, 14, nstart = 25)” and visualized using the function “fviz_cluster(optimal, data = X)”. Further sub-groups in each cluster were identified using the “FactorMineR” package which computes the hierarchical clustering using the principal components of the individual dataset.

Results

Scoring and abundance of SSPs bands across genotypes

Based on scoring, a total of 44 bands were identified which were all polymorphic, their molecular weights were in the range of 10 kDa (kilodalton) to 170 kDa. The least appeared protein bands were found in some genotypes, for instance, a band of 170 kDa was detected in two wild types species only and was absent in all cultivated and wild type chickpea. Similarly, another band of 160 kDa was present in 23 wild type and two cultivated accessions only. However, five bands of sizes 140, 150, 120, 16, and 12 kDa were present in < 10% of the accessions. Another band of 11 kDa was present in 33 wild types but was absent in all cultivated types. The band size of 126, and 145 kDa were present in 46 accessions. Moreover, the band of 130 kDa was present in 58 accessions. Two bands sizes of 10 and 15 kDa were present in 56 accessions. Likewise, bands of sizes 14 and 17 kDa were present in 70 accessions each.

The 11 protein bands of sizes 115, 87, 75, 62, 60, 45, 48, 28, 27, 24, and 20 kDa were detected in up to 200 accessions. Likewise, peptides of sizes 113, 100, 98, 85, 70, 58, 55, 50, 30, 25, 21, and 13 kDa were detected in up to 300 genotypes of cultivated as well as wild type. Finally, the most abundant peptides were of the sizes 35 and 37 kDa detected in 398 and 395 accessions respectively. Additionally, there were five peptides of sizes 66, 42, 37, 35, and 33 kDa were detected in more than 300 genotypes. The results are described in detail for each band and accession in Table 1, and represented graphically in Fig. 1. A representation of the electrophoregram and names of the accessions is provided in Fig. 2.

Fig. 1
figure 1

Number of bands of various sizes in cultivated (light blue) and wild (navy blue) accessions of the chickpea germplasm

Full size image
Fig. 2
figure 2

A representation of electrophoregram; a serial number of accessions from Table S1, b protein banding patterns of the accessions on SDS-PAGE and Protein ladder (Broad range, Bio-Rad) with molecular weight (Dalton), c technical names of the accessions, d clusters of the respective accessions and e scoring of the bands according to which clusters were determined. Unique bands (present in only one accession) are highlighted in deep pink

Full size image
Table 1 Abundance and percentage (abundance) of 44 SSP bands in 436 chickpea cultivated and wild genotypes
Full size table

Stratification and cluster analysis

We employed the three most commonly used methods viz. “elbow method” [30], “silhouette method” [31], and “gap statistic” [32] to determine the appropriate number of clusters in our germplasm. Using the “elbow method”, no optimal clusters could be determined in our dataset (Fig S1). The silhouette method indicated the presence of three possible clusters in our dataset (Fig S2). However, when the “silhouette method” model to cluster construction was fitted to three clusters, many accessions in clusters 1 and 3 were found concentrated together. As a result, we failed to differentiate among those accessions and hence, those accessions remained unrecognizable (Fig S3). Employing the “gap statistic” method [32], based on the presence or absence of protein bands, all genotypes could be divided into 14 clusters (Fig. 3) where the number of genotypes in a single cluster varied from eight (cluster 7) to 58 (cluster 12) (Fig. 4). However, an adequate amount of overlapping was also present in different clusters (Fig S4).

Fig. 3
figure 3

Determination of optimal number of clusters based on “gap statistic”

Full size image
Fig. 4
figure 4

Number of clusters (inner circle) and the number of accessions they contained (number in brackets), the outer circle indicates the number of sub-clusters in each cluster 

Full size image

In addition to the main 14 clusters, the genotypes within each cluster were further sub-divided into several sub-clusters (Table S2-S15, Fig S5-S18). The number of sub-clusters within each cluster varied from three (clusters 2, 4, 8, 10, and 12) to six (cluster 3). Five subgroups were detected in cluster 9 whereas the rest of the clusters (1, 5, 6, 7, 11, 13, and 14) contained four sub-clusters (Fig S2-S15). Cluster 1 contained 48 genotypes originated from 16 different countries where the major contribution is from India (12 genotypes) and Turkey (13 genotypes) (Table S2). In addition, a total of 15 were advanced cultivars, three were breeder’s line, 21 were traditional cultivars and four were wild types. Except for the wild types (Cicer reticulatum), all were C. arietinum. The second, fourth, and seventh clusters contained 10, nine, and eight genotypes, correspondingly, all of which were elite cultivars originating from Pakistan (Tables S3, S5, and S8). The third cluster carried 31 genotypes originating from 10 different countries. Nine of them were advanced and three were traditional cultivars. The rest of the 19 genotypes were wild types. The wild types belonged to C. echinospermum (one genotype), C. reticulatum (8 genotypes), and C. judaicum (10 genotypes) (Table S4). Cluster 5 contained 34 wild types that originated pre-dominantly from Turkey (Table S5). There were 40 genotypes grouped in cluster 6 originating from Turkey (27 genotypes), Russia (eight genotypes), India (one genotype), Spain (one genotype), and Syria (one genotype) whereas the origin of two genotypes remained unknown (Table S7). In this cluster, the Syrian and other 26 Turkish genotypes were wild types while the Russian, Indian, Spanish and one Turkish genotype were cultivated types. The eighth cluster was comprised of nine cultivated types originating from different countries (Table S9). There were 37 genotypes in cluster 9 of cultivated type (advanced or traditional cultivars or elite lines) originating from many different countries including 14 Syrian genotypes (Table S10). The 10th cluster carried 51 genotypes originating from 14 different countries where all except one were cultivated types (Table S11). Cluster 11 contained 34 genotypes, among those all were cultivated species and most of them (19) have Indian origin, and three genotypes were from Turkey. (Table S12). In cluster 12, 50% of the genotypes originated from India (15 genotypes) and Turkey (14 genotypes) (Table S13). In the 13th cluster, there were 35 genotypes where 26 were cultivated types and 9 were wild type belonging to 15 different countries (Table S14). Among them, the wild types originated from Israel (two genotypes), Morocco, Russia, Syria, and Turkey (four genotypes). The 14th cluster carried 32 genotypes originating from Turkey (except one genotype that originated from Israel), all of which were wild types (Table S15).

Discussion

Among old techniques, one of the reliable and economic techniques to study population genetics is the SDS-PAGE of seed storage proteins because of little or no influence of environmental fluctuations on them [33]. Earlier, the characterization of seeds storage proteins are part of the basic and applied studies of any crop plant research [34]. Plant populations of chickpea [25, 35] and cowpea [34] have successfully been differentiated based on seeds storage proteins. This differentiation is necessary to select superior but diverse parents for future breeding [23].

Polymorphism of protein bands in context to their potential functions reported in previous studies

Seed storage proteins are proteins that accumulate significantly in the developing seed. The majority of seeds contain one to many groups of proteins, which serves as a source of amino acids, metal ions, nitrogen, carbon, and sulfur. These proteins are rapidly mobilized during seed germination [36]. In chickpea, globulins are the dominant storage proteins and account for 50–90% of seed protein followed by glutelins (18.10%), albumins (12.0%), prolamins (2.8%), and others [37]. These groups of seed storage proteins have been characterized and assigned a specific range of molecular weights which helps to locate the presence and absence of these proteins on a Coomassie stained SDS-PAGE gel. For example, according to previous reports the protein bands of 10, 11, and 12 kDa correspond to albumin (2 S) proteins [38]. Likewise, bands of 15, 18, 33, 35, 37, and 50 kDa correspond to vicilin (7 S) proteins [39]. Similarly, peptides of sizes 24, 25, and 42 belong to legumin (11 S) proteins whereas those of sizes 55 and 58 kDa belong to glutelin proteins [40, 41]. Besides, we have identified 44 protein bands across all 436 chickpea genotypes which could belong to albumin, vicilin, legumin, and glutelin proteins.

Globulins

Globulins are the most abundant form of seed storage proteins and were monomorphic as these were found in all genotypes. Vicilins together with legumins form globulins fraction of the legume seeds representing 70% of SSPs [42]. Legumins constitute 97% of globulins and are considered a major impetus during germination and seedling growth in chickpea [43]. Our results validate these findings as globulins were present in all accessions and were most abundant. All accessions carried at least one of the six vicilin (7 S) proteins (Fig. 5a). A total of 17 accessions exhibited one vicilin protein. Likewise, another 17 accessions showed all the six vicillin protein bands. Forty-three accessions had at least two vicilin proteins whereas 139 and 150 accessions carried three and four vicillin proteins, respectively. Among all 66 accessions carried five vicilin proteins. Legumin (11 S) proteins were present in 399 of the total 436 accessions whereas 57 accessions carried legumins of all sizes (Fig. 5b). In addition, 112 and 230 accessions correspondingly carried two legumins.

Fig. 5
figure 5

Fractions of a Vicillin, b legumin, c glutelins, and d Albuminseed proteins

Full size image

Glutelins

Glutelins contain higher levels of methionine and cysteine [44]. These are, however, not very well characterized [45]. Among all genotypes, 229 accessions showed one or two bands representing Glutelins (Fig. 5c). Glutelins were found more polymorphic and can be used as markers in plant breeding.

Albumins

Sulfur amino acids constitute a large proportion of Albumin proteins. Our results described that two, out of 3 peptides of Albumin (2 S) proteins were present in 80 accessions out of which 61 were wild and 19 were cultivated species (Fig. 5d), however, 10 wild type accessions from Turkish origin carried all three peptides. All 19 cultivated types carried either 10 (9 accessions from Pakistan) or 12 kDa (nine Russian and one Tajik accession) peptides of Albumin. Albumins are known to influence certain physiological functions including activities of lipoxygenases, glycosidases, or proteases involved in the degradation of storage proteins. Lipoxygenases influenced longevity in wheat [46]. Other Albumins like lectins are reported to play roles in plant’s defense against various stresses [39, 47]. Furthermore, our results need to be evaluated for their potential role in stress tolerance and use Albumin protein bands as a protein marker for germplasm screening against plant stresses.

Cluster analysis

Sometimes, agronomic characteristics and ecotypes are not sufficient to differentiate between various accessions of the same plant species, nevertheless, cluster analysis based on the SDS-PAGE scoring of seed storage proteins (SSPs) has differentiated among Vigna radiata and Vigna mungo accessions [24]. Clustering finds subgroups of observations within a dataset based on similarity, were observations in the same group tend to be similar and observations in different groups tend to be dissimilar. K-means, clustering method splits the data into k groups through the k-mean algorithm which is a local search procedure but suffers from serious drawbacks [48]. In the case of chickpea, cluster analysis has been proved as a powerful tool to distinguish between a set of germplasm [35, 49]. The reported studies used only a few accessions, eight by [25]), 34 by [35], and 118 by [49] whereas we have used 436 accessions in this investigation. No studies are available with respect to seed storage protein markers based clustering for this size of the dataset. We used three clustering methods i.e. elbow method, silhouette method, and gap statistics to understand the latent structure of the genotypes. Our results established that the gap statistics method is the most appropriate to determine the number of clusters in a large dataset as described earlier in the text.

Previous studies showed that 20 clusters were identified in 118 genotypes of exotic and local germplasm, where the maximum number of genotypes in one single cluster was 47, mainly from Syria and the USA [49]. Three clusters out of eight Kabuli chickpea mutants/hybrids and four clusters out of a set of 34 genotypes were detected in previous studies by [25, 35] respectively. Moreover, no sub-clusters were reported in the aforementioned studies.

We have identified 14 main clusters as well as each cluster harboring three to six sub-clusters in this study. Our results demonstrated that 27 Pakistani genotypes clustered separately in cluster numbers 2, 4, and 7 and further into 3, 3, and 4 subclusters respectively (Fig S7, S9, and S12). Nevertheless, all the clusters (2, 4 and 7) containing Pakistani germplasm did not contain any exotic or wild genotypes implying the possible use of few progenitors during their domestication and inbreeding.

Based on our findings it is assumed that one method to increase the diversity in our germplasm which is a prerequisite for getting wheat varieties with desirable traits; is to include wild as well as exotic genotypes in future breeding [23]. For instance, in the case of wheat, the reason behind that the existence of a relatively small gene pool of wheat cultivars in Pakistan [50] is the use of recurrent selection which further created the omnipresent problem of rust epidemics. The same could happen in case of chickpea.

Hence, we propose a three-way cross of distinctly related genotypes, for example by including genotypes from cluster 2 (for example genotype number 188 viz. CM2000) and 7 (for example genotype number 197 viz. PakK01219) and any of the wild or exotic accessions from (for example genotype number 36 WALSGOTT-3 from Egypt) to increase diversity as well as selection gain in Pakistani chickpea breeding program.

It has been demonstrated before [25] that protein markers of certain sizes (KSSP-100, KSSP-93, and KSSP-64) can be used for the identification of CM-98/99, ILC-195 and CM-2000 genotypes, respectively. We propose here that polymorphic protein bands ranging in size from 10, 11, 12, 14, and 15 kDa and 120,126, 140, 145, 150, 160, and 170 kDa can be used to detect the introgression of blood from wild species into the hybrids produced by crossing wild species harboring desired traits like stress tolerance with cultivated chickpea genotype. We have observed that proteins markers of 10, 126,140, and 160 kDa in size were mostly present in wild type as compared to cultivated types. Specifically, we found that two Pakistani-approved cultivars showed a protein marker of 160 kDa named CH-2016 and Pb-2008. As the name indicated, CH-2016 is a recent variety from NIAB, Pakistan with excellent yield potential and high demand among chickpea growers. Pb-2008 was also a successful variety in 2008 and yet is being cultivated in some areas where it is specifically adapted. We propose that the yield advantage of these two genotypes and their wider adaptability could be due to the incorporation of genetic background from wild type in their ancestry and can be used as a marker in chickpea breeding. This need to be further investigated which is not within the scope of this study.

Conclusion

At large, cluster analysis of 436 chickpea cultivated and wild type genotypes on the basis of SSPs profiling could be differentiated into 14 main clusters with three to six sub-clusters in each main cluster. There were three main clusters of 27 Pakistani genotypes only. Pakistani germplasm showed rare protein bands which were not otherwise present in any other genotype including both cultivated and wild types. Further, genotypes from the three different main clusters could be crossed with other distinct genotypes in the study to develop hybrids to hasten the breeding high yielding cultivars with wider adaptability. In addition, the very high and very low sized protein bands can be used as possible markers of exotic genetic background in the cultivated germplasm, identification of hybrids and to stress tolerant genotypes, for instance albumins with their known role in stress tolerance can be used as marker in chickpea breeding. In the end, it is recommended that SDS-PAGE of SSPs is an important and cost effective technique and can be used successfully to identify hybrids among distantly related parents, measure genetic diversity across genotypes grown under a range of global environments with authenticity and it is worth to be established in laboratories where cost matters, due to its cost effectiveness.