Genetic Diversity and Population Structure of Two Faba Bean Mutant Populations Based on AFLP Markers

Faba bean is the fourth most produced cool-season legume worldwide, with the total production of dry grains up to 4.92 million tons from 2.51 million ha-harvested areas in 2018 (FAOSTAT 2018). Faba bean has potency as a future protein supply for the human diet because of its chief protein source, fiber and other non-nutrient compounds (Multari et al., 2015). Oliveira et al., (2016) stated that faba bean also has an essential role in soil fertility and nitrogen fixation. Despite its importance, Nedumaran et al., (2015) reported that the growth rate of faba bean’s yield was the lowest among legume crops. Therefore, the development of new high-yielding faba bean varieties is needed to boost faba bean production. The development of high-yielding varieties depends on the availability of genetic diversity, which can be enriched by induced mutagenesis. The genetic diversity generated by induced mutagenesis has been reported in many studies that included faba bean (Nair and Mehta 2014; Shekar and Pushpendra 2017; Khursheed and Khan 2017; Singh and Sadhukhan 2019). However, most of these studies were based on morphology-based diversity. Plant breeders widely use morphology-based markers to determine genetic diversity because of their simplicity and low cost. However, only a few traits can be tested. Environmental factors also affected most economically important traits, making it difficult to test using morphological markers.
       
Molecular or DNA-based markers are abundant and are not influenced by the environment. Therefore, DNA-based markers serve as a powerful tool for characterizing genetic diversity in various crop species, including faba bean (Alghamdi et al., 2012a). One of the best molecular markers for studying genetic diversity among the faba bean populations is amplified fragment length polymorphism, AFLP (Zeid et al., 2003; Zong et al., 2010, Alghamdi et al., 2014; Nurmansyah et al., 2020a). The AFLP-based molecular marker significant advantage is the high reproducibility or the capacity to generate high polymorphism bands, which is useful for evaluating genetic diversity.        
       
Although induced mutagenesis can rapidly generate genetic diversity, every genotype responds differently to different mutagen treatments in inducing genetic diversity. Therefore, this study compares and tests the genetic diversity of the two genotypes of faba bean treated with two mutagens based on AFLP markers in the M₂ generation. The study is crucial to select the most diverse population and help reduce population size for further generations. The two genotypes were Hassawi 2, representing a faba bean landrace cultivar from Saudi Arabia and ILB4347 represents an inbred line from ICARDA. Hassawi 2 is reported drought stress tolerance and well adapted in arid conditions of Saudi Arabia (Migdadi et al., 2016), while ILB4347 is reported to have salt stress tolerance and resistance to broomrape (Alzahrani et al., 2019; Rubiales et al., 2016). The two mutagens were gamma radiation and diethyl sulfate (DES). We investigated the population structure within mutant populations to assess population change in the two genotypes, landrace and inbred populations.

Plant Materials and DNA extraction
 
5271 M₂ seeds generated from the M₁ generation of two faba bean genotypes, Hassawi 2 and ILB4347, were used in this study. The original seeds of Hassawi 2 were obtained from the Legume Research Group, Department of Plant Production, College of Food and Agriculture Sciences, King Saud University, and the seeds of ILB4347 were obtained from ICARDA. The M₁ seeds were produced by exposing separate 120 dry seed samples of each genotype to 25 and 50 Gy at a dose rate of 15.48 Gy/min of gamma radiation using a ⁶⁰Co (Cobalt 60) gamma source (Gamma Chamber 900, Nordion, Canada) under ambient conditions at the Central Lab, College of Science, King Saud University, and three concentrations of DES 0.01%, 0.02% and 0.05% solution. The dry seeds were soaked in DES solution for 12 h at room temperature and then were thoroughly washed under running water three times (Nurmansyah et al., 2018).
       
Two hundred eighty of 4102 mutant plants, representing 280 mutant families, were chosen from the mutant populations. Selection criteria of selected mutants were based on survival rates at the M₁ and M₂ generations, followed by phenotypic mutants described by our previous study, including determinate growth habit (Nurmansyah et al., 2019), flower, leaf and seed character mutants (Nurmansyah et al., 2020b) and quantitative characters such as the number of branches, pod length, number of seeds per pod and 100 seed weight. Leaves of 280 M₂ mutant plants derived from 140 mutant plants of Hassawi 2 and 140 mutant plants of ILB 4347 and two parental lines were collected for DNA isolation. As described by Alghamdi et al., (2012b), a modified SDS protocol was used for DNA isolation.
 
AFLP procedure
 
The AFLP plant mapping protocol from (ABI) Applied Bio-systems (Waltham, MA, USA) was used in this study with some modifications as described by Nurmansyah et al., (2020a). EcoRI/MseI and T4DNA-ligase enzymes used for digestion and ligation of DNA samples. The AFLP adapter/primer name and sequence used are shown in Table 1. The pre-selective amplification was performed using primers that matched the adapter sequence and had one additional ‘selective’ base (EcoRI + G and MseI + C). Selective amplification used the diluted pre-selective amplification product as the new template. This reaction’s primers had the same sequence as the pre-selective primers with two and three other selective bases in EcoRI and MseI primers. Eleven AFLP primer combinations were used to estimate genetic diversity among faba bean mutant lines. AFLP fragments were performed using GeneMapper Analysis Software v3.7 (ABI) and the data assembled in a binary format, with 1 representing allele presence and 0 for absence. The threshold for allele calling is set at 100 relative fluorescence units (rfu). We assigned peaks at 100 rfu or higher assigned as one and those with lower 100 rfu a zero value. We performed fragment analysis for allele sizes between 100 and 500 bp, as recommended by Paris et al., (2010).
 

       
The total number of alleles, the number of bands, the average number of bands per sample, and the average number of bands per allele recorded for each primer. Polymorphism information content (PIC) estimated according to Botstein et al., (1980) and discrimination power (DP) calculated by dividing the number of polymorphic alleles amplified for each primer by the total number of polymorphic alleles recorded. Dendrogram using the Jaccard similarity coefficient and the unweighted pair group method with arithmetic average (UPGMA) constructed using Paleontological Statistics (PAST) v. 3.20 program (Hammer et al., 2001).
 
Genetic diversity parameters and population structure
 
The presence/absence binary matrix was used for further analysis with GenAlEx 6.503 complement for MSExcel (Peakall and Smouse 2012). The total number of different alleles (Na), number of effective alleles (Ne), Shannon’s information index (I), expected heterozygosity (He), percentage of polymorphic loci (% P) and private alleles per population estimated using this approach. Pair-wise, Jaccard genetic similarity using Paleontological Statistics (PAST) v. 3.20 program used to assess genetic diversity among the 280 mutant and control plants. Analysis of molecular variance (AMOVA) was performed using GenAlEx 6.503 with 999 permutations. Using STRUCTURE 2.3.4 program, the population structure was tested and supported by Principal Coordinates Analysis (PCoA) using the PAST program based on Jaccard’s similarity index.

Assessment of genetic diversity based on AFLP-molecular marker
 
The characteristics of 11 AFLP primer combinations are presented in Table 2. 2083 alleles were generated from 282 samples (280 mutant plants and two control plants). The number of alleles ranged from 97 for primer combination (EcoRI/MseI) TT/CTT to 305 for primer combination TG/CTT. The 11 primer combinations generated 89,820 bands with an average of 8,165.45 bands per primer combination. The average bands recorded across the samples were 28.96 bands per sample, ranged from 15.48 to 60.63 bands. Simultaneously, the average bands per allele ranged from 31.57 to 56.06 with an average of 42.70 bands.
 

       
The analysis of genetic diversity parameters is presented in Table 3. The analysis was divided based on the initial genotype and the combination between the initial genotype and treatment, gamma radiation (physical mutagen) and DES (chemical mutagen). The genetic diversity based on the initial genotype revealed that the mutant population of ILB 4347 was more diverse than the mutant population of Hassawi 2, as shown by the higher value in the genetic diversity parameters. Based on the combination between the initial genotype and treatment, four subpopulations were analyzed and revealed that the highest genetic diversity was found in the mutant population of ILB 4347 induced by DES. The lowest genetic diversity was recorded in the mutant population of Hassawi 2 induced by DES. Genetic diversity comparison among the mutant populations revealed that each genotype responds differently to different treatments in inducing genetic diversity.
 

       
Based on genetic diversity parameters, treatments with gamma radiation generated the maximum genetic diversity in Hassawi 2, while in ILB4347 mutant populations, DES generated higher genetic diversity than the gamma radiation. It showed that gamma radiation is more efficient to enrich genetic diversity than DES in Hassawi 2. In ILB4347, DES is more efficient to enrich genetic diversity than gamma radiation. Therefore, for further generation, gamma radiation-induced and DES-induced populations should be selected in Hassawi 2 and ILB4347, respectively.
       
The AMOVA analyses showed that the maximum genetic variation is found within the populations. These results align with other faba bean studies (Wang et al., 2012; G  l et al., 2017; El-Esawi, 2017). The two-level analysis of AMOVA based on the initial genotype showed that 94% of the variation was concentrated within genotypes, and the variations between the initial genotype population contributed to the remaining 6% (Table 4). All the mutant plants from both genotypes subjected to the same treatments may have equal chances to contribute to both genotypes. A similar likelihood of inducing diversity in both genotypes was supported by a relatively similar value seen in the pair-wise Jaccard’s genetic similarity index between populations. The similarity index within Hassawi 2 population ranged from 0.05 to 0.65, whereas the similarity index within the ILB4347 population ranged from 0.05 to 0.67. The 6% variation was observed, possibly because of the different genetic makeup between the genotypes or random mutation events.
 

 
Cluster analysis
 
Cluster analysis of populations based on Nei’s genetic distance revealed that two clusters could be formed based on the initial genotype if the populations were grouped by genotype-treatment combinations (Fig 1). Fig 1A shows two clusters formed with a genetic similarity value of 0.54 (50% of genetic similarity). Cluster one and two represent ILB 4347 and Hassawi 2 populations, respectively. The PCoA analysis also clearly divided the populations into two clusters through the first and second principal coordinates that accounted for 73.6% of the total variations (Fig 1B). Cluster analysis-separated faba bean mutant populations based on genotypic origin. The population could be divided into two clusters; cluster one for the ILB 4347 population and cluster two for the Hassawi 2 population. This grouping was based on a comparison of four mutant populations based on combinations between genotypes and treatments. This grouping revealed the pedigree of the populations. Fig 2B shows a clear separation of the clusters, but the distance between the mutant populations in genotype Hassawi 2 is higher than that in ILB4347 mutant populations. Based on the AMOVA analysis, in Table 5 and 6, 10% variation is observed between the populations induced by gamma radiation and DES in Hassawi 2, whereas only 3% variation was observed between the treatments ILB4347 mutant populations.
 

 

 

 

 
Population structure
 
The first panel of the population structure examined the population structure within the Hassawi 2 mutant population. Evanno’s DK peaked at K = 3 (Fig 2A), showing that mutant plants within the Hassawi 2 population could be divided into three clusters. Fig 2B shows the results of STRUCTURE on Hassawi 2 mutant populations. Red lines represent cluster one and green and blue lines represent clusters two and three, respectively. The second panel of population structure tested the ILB 4347 mutant population structure. Evanno’s DK value peaked at K = 2 (Fig 2C), showing that the mutant plants can be divided into only two clusters. Fig 2D shows the STRUCTURE analysis of ILB 4347 mutant populations categorized as major and minor clusters. A significant cluster or cluster two is represented in green and a minor cluster or cluster one is represented in red.
       
The population structure analyses showed a different population structure between Hassawi 2 and ILB4347 mutant populations. Hassawi 2 mutant populations were clustered into three groups, whereas ILB4347 mutant populations were clustered into two groups based on the best Evanno’s DK inferred by STRUCTURE HARVESTER software. The different number of clusters between the two mutant populations shows their breeding history. Sim et al., (2011) revealed that population structure data positively correlated with the population’s breeding history. Based on this knowledge, distinct patterns of population structures between Hassawi 2 and ILB4347 mutant populations could be because of differences in their breeding history. Hassawi 2 is a landrace cultivar and it is probably genetically more diverse than the inbred line (ILB4347). Therefore, the number of gene pools influencing the population structure was higher in the Hassawi 2 population than in the ILB4347 population.

This work is supported by the NSTIP strategic technologies program number (11-AGR1861-02) in Saudi Arabia.