Association of Seed Amylose Content with Agronomic Traits and GBSS1a Haplotypes in Common Bean Landraces

Common bean seeds are a prominent source of complex carbohydrates (50-60%), proteins (20-25%) and minerals (Basavaraja et al., 2021; Rasool et al., 2019). Starch represents a major nutritional constituent of carbohydrates in the human diet and is composed of amylose and amylopectin (Tayade et al., 2019). Amylose is a linear structure linked by a-1,4-glycosidic bonds and is synthesized by granule-bound starch synthases (GBSSs), whereas amylopectin is a branched structure linked by both α-1,4-glycosidic and α-1,6-glycosidic bonds and is synthesized by soluble starch synthases (SSs), branching enzymes (BEs) and debranching enzymes (DBEs) (Dupuis et al., 2014). Given their distinct structures, the amylose-to-amylopectin ratio influences the properties of starch (Dupuis et al., 2014). Among these two starch components, the content of amylose was known to be highly correlated with the quality of starchy foods, as it determined the pasting and gel-texture properties of starch during cooking (Punia et al., 2020). Therefore, evaluating the amylose content of common bean accessions can provide a strong foundation for improving the quality of bean-derived food products.

Despite the importance of common bean in the human diet, most studies on starch and amylose have been conducted on cereal crops and only a few on legumes (Tayade et al., 2019). Moreover, among the studies conducted on legume starch, most were performed on pea (Pisum sativum L.). For example, Hybl et al., (2001) assessed variation in starch and amylose content in 402 pea accessions; Jha et al., (2015) examined allelic variation in genes associated with amylose and total starch contents of 169 pea accessions; and Carpenter et al., (2017) performed an association analysis on 25 starch metabolic genes and amylose content of 92 pea accessions. Among the studies conducted on common bean, Rivera et al., (2018) evaluated variations in the seed starch and amylose content of 202 common bean accessions, while Isono et al., (2003) and Matsui et al., (2003) reported that GBSS1a and GBSS1b genes were mainly involved in the synthesis of amylose in storage starches in seeds and amylose in transitory starches in leaves, respectively. However, the relationship between GBSS1a haplotype and seed amylose content in common bean has not been investigated.

Therefore, the aims is to (i) evaluate the diversity of seed amylose content in common bean; (ii) elucidate the relationship between amylose content and agronomic traits; and (iii) analyze the GBSS1a gene haplotypes, which are likely to be associated with amylose content. The results improve the understanding of the genetics of seed amylose content in common bean. Additionally, correlation analysis of amylose content and agronomic traits provides sophisticated information about the breeding material that could be used for improving the starch composition of common bean.

Plant materials
 
A total of 300 common bean landraces originating from six countries, including Korea (44), China (43), Georgia (64), Ukraine (38), Bulgaria (66) and El Salvador (45), were obtained from the National Agrobiodiversity Center (NAC) of the Rural Development Administration (RDA), Republic of Korea (http://genebank.rda.go.kr). The common beans were sown in plug trays on April 9, 2020 and the seedlings were transplanted into the experimental field of NAC on April 29, 2020. The planting distances were 110 cm between the rows and 40 cm between each other in a row. Each accession consisted of 10 plants and twenty agronomic traits were recorded in all of the plants during various stages of their growth. Leaf samples were collected from ten 14-day-old plants of each accession and used for DNA extraction. At maturity, the pods were harvested and threshed before determining the amylose content in the seeds.
 
Agronomic trait characterization
 
Twenty agronomic traits (9 qualitative and 11 quantitative traits) were characterized according to the NAC descriptors (Table S1).



Qualitative traits were recorded as representative values of 10 plants per accession and quantitative traits were recorded as their average. The average values of measured pod length (PL), pod width (PW) and seed number per pod (SNPP) were taken from 10 pods. The seed area (SA), perimeter (SP), length (SL), width (SW) and length/width ratio (L/W) of 20 seeds per accession were measured using ImageJ software (Schneider et al., 2012).
 
Determination of seed amylose and starch contents
 
The experimental design was completely randomized and incorporated biological duplicates. Seed starch content (mg/100 mg flour) was determined using the rapid total starch (RTS) method (Megazyme International Ireland, Ltd., Ireland). Seed amylose content (mg/100 mg flour) was determined with the colorimetric method using I2/KI, as described by Juliano (1971). The percent amylose content of seeds was estimated based on their amylose and starch contents (mg/100 mg flour), as follows:

 

 
 GBSS1a haplotype analysis
 
Genomic DNA was extracted from 20 mg freeze-dried leaves of each accession using the Gentra Puregene Tissue Kit (QIAGEN, Inc., USA). Six primer sets flanking different regions of the GBSS1a reference gene sequence (Phvul.001G082500; source, P. vulgaris cv. ‘G19833,), from the 5' untranslated region (5'UTR) to 3'UTR (3,919 bp), were designed using the NCBI primer designing tool (https://www.ncbi.nlm.nih.gov/tools/primer-blast/) (Table S2).



PCR was performed in a 20 mL reaction volume containing 1 mL genomic DNA (50 ng/mL), 1 mL each of forward and reverse primers (10 pmoles/mL), 10 mL of 2 × Direct PCR premix (Inclone Biotech, Republic of Korea) and 8 mL distilled water. PCR was performed on the SimpliAmp Thermal Cycler (Thermo Fisher Scientific, USA) under the following conditions: initial denaturation at 95°C for 5 min, followed by 35 cycles at 95°C for 15 s, 58°C for 30 s and 72°C for 1 min, followed by a final extension at 72°C for 5 min. Then, 20 mL of the PCR product of each reaction was mixed with 4 mL Dyne LoadingSTAR (Dyne Bio, Republic of Korea) and samples were electrophoresed on a 1% agarose gel in 0.5´ TBE buffer at 135 V for 40 min. DNA-containing gel slices were excised using a razor blade under a UV transilluminator and DNA was purified using the QIAquick Gel Extraction Kit (QIAGEN, Inc., USA). The purified DNAs were sequenced by GenoTech, Corp. (Daejeon, Republic of Korea). Multiple DNA sequences were aligned with the reference gene sequence using BioEdit version 7.0.5.3 (Hall, 1999) to identify polymorphisms.
 
Statistical analysis
 
The R software (version 4.0.2) was used for all statistical analyses. Analysis of variance (ANOVA) was performed to compare the amylose content of various genotypes. The Pearson correlation coefficient was determined to test the correlation of amylose content with various quantitative traits. K-means clustering analysis was performed using 11 quantitative agronomic traits.

Variation in agronomic traits
 
Nine qualitative traits were evaluated among the 300 bean landraces and diverse phenotypes were observed (Fig 1).



Approximately 50% of the germplasm showed an indeterminate upright growth habit, while the remaining 50% showed determinate and indeterminate bush growth habits. The banner and wings of flowers showed both solid and patterned colors; however, only a few accessions displayed patterned colors in the wings. Pods showed more diverse colors at maturity than at harvest stage. Most bean accessions produced slightly curved or straight pods containing weak- or medium-brilliance seeds and the seeds showed eight different colors and five different shapes. All 11 quantitative traits examined in this study showed wide variation between the bean landraces (Table 1).



DTF (days to flowering, 53.74±5.22 days) and DTM (days to maturity, 87.42±11.47 days) were approximately 1 month apart. The average PL and PW were 11.51±2.44 cm and 12.09±2.88 mm, respectively. Among the seed-related traits, SA (79.00±21.81 mm²) and 100SW (100-seed weight, 38.04±12.86 g) showed the greatest variation. The K-means clustering results revealed three clusters (Fig S1).



The clusters 1, 2 and 3 consisted of 132, 75 and 93 accessions, respectively. Average values of SA, SP, SL, SW, 100SW and PW increased in the order of 1, 2 and 3, whereas average values of SNPP decreased, indicating that these traits were the primary clustering determinants. Also, cluster 3 had higher DTF and DTM values than clusters 1 and 2. The average values of PL and L/W increased in the order of cluster 2, 3 and 1.
 
Variation in seed amylose and starch contents
 
The starch, amylose and percent amylose contents of bean seeds varied considerably between the landraces. The average values and ranges of starch, amylose and percent amylose contents were 32.21±3.32 mg/100 mg flour (range, 22.20-43.36 mg/100 mg flour), 15.51±1.67 mg/100 mg flour (range, 10.03-19.70 mg/100 mg flour) and 48.24%±3.19% (range, 39.95-63.57%), respectively (Table 2).



Previous studies on common bean cultivars revealed variation in seed starch and amylose contents, ranging from 15.4 to 60 mg/100 mg flour and 10.2% to 51.1% amylose content (Keskin et al., 2022; Tayade et al., 2019). Whereas the range of starch content of 300 common bean landraces investigated in the current study (22.20-43.36 mg/100 mg flour) was within that reported previously, the range of amylose content of these landraces (39.95-63.57%) was not completely consistent with previous data. Among the 300 landraces examined, the amylose content of 22 landraces was more than 52%. This extreme variation in seed amylose content observed between the 300 landraces could be explained by the diverse geographical origins of the landraces (Kanwar and Mehta, 2018; Murube et al., 2021). Moreover, common bean starch contains the highest amylose content compared with other legume starches (Punia et al., 2020); therefore, finding bean accessions with high amylose content is not difficult.
 
GBSS1a haplotype analysis
 
Relative to the GBSS1a reference gene sequence of ‘G19833’, the common bean germplasm investigated in this study showed two non-synonymous single nucleotide polymorphisms (SNPs) in exon 10 of the GBSS1a gene at 2,748 and 2,878 bp positions (Fig 2).



The SNPs at 2,748 and 2,878 bp were predicted to cause tyrosine (Y, TAT) to asparagine (N, AAT) and proline (P, CCA) to glutamine (Q, CAA) substitutions, respectively. Three different SNP haplotypes were identified in the germplasm, which were named according to the combination of amino acids: YP haplotype, which was identical to the reference sequence and YQ and NP haplotypes, which carried one amino acid substitution relative to the reference sequence (Fig 2). Notably, the NQ haplotype, carrying two amino acid substitutions relative to the reference, was not found among the 300 common bean landraces investigated in this study.
 
Correlation analysis between amylose content and agronomic traits
 
We examined the correlation between seed amylose content and 11 quantitative agronomic traits by determining the Pearson correlation coefficient at p<0.05 (Fig 3).



The results showed that amylose content was positively correlated with PW, SW, SA, 100SW, SP and SL. Purwanti et al., (2019) reported a significant positive correlation between amylose content and seed size in lablab bean (Dolichos lablab L.) seeds, consistent with the current study on common bean seeds. On the other hand, we also found that SNPP showed a negative correlation with amylose content and seed size, implying that an increase in SNPP is accompanied by a decrease in seed size, resulting in lower amylose content. Furthermore, analysis of the correlation between amylose content and seed weight in previous studies produced different results, depending on the plant species or germplasm under investigation. For example, Rivera et al., (2018) reported no correlation between amylose content and 100-seed weight in a Spanish core collection of common bean accessions, whereas Tahir et al., (2011) reported a negative correlation between amylose content and 1,000-seed weight in lentil (Lens culinaris Medik.). In contrast to these studies, we identified a positive correlation between amylose content and 100SW in a collection of 300 common bean landraces.
 
Correlation analysis between amylose content and GBSS1a Haplotype
 
Common bean landraces with the YP, YQ and NP haplotypes of the GBSS1a gene contained 48.31%±3.52%, 47.57%±2.76% and 49.37%±3.08% amylose, respectively. Landraces with the NP haplotype showed higher amylose content than those with YQ haplotype (p<0.05) (Table 3).



However, this correlation could not fully explain the tremendous amount of variation in seed amylose content in bean germplasm, demonstrating amylose content is a quantitative trait influenced by numerous genes.

In this study, we evaluated the range of diversity in the seed amylose content, agronomic traits and GBSS1a gene haplotypes of 300 common bean landraces and examined potential correlations between these variables. The bean landraces showed a wide variation in amylose content. Although this diversity in amylose content was not fully explained by its correlation with agronomic traits or GBSS1a gene haplotypes, this study contributes toward improving the amount of information currently available on starch content and composition in legumes. Haplotype analysis of other starch synthesis genes, such as BEs and DBEs, in common bean would pave the way for further advancements in research on legume starch.
 

This research was funded by the Research Program for Agricultural Science and Technology Development, National Institute of Agricultural Sciences, Rural Development Administration, Republic of Korea, grant number [PJ01424201].

None