Twenty improved durum wheat varieties of Ethiopian origin were included in the present study. Their local names, pedigrees, origin and adaptation areas are given in Table 1. The seed samples grown in Ethiopia in 2017 were used in this experiment. All analyses were conducted in the same year in the Department of Crop Science and Biotechnology, Jeonbuk National University, Jeonju, South Korea [11].

One hundred milligrams of finely ground wheat flour was used to sequentially extract the four major protein fractions by the Osborne procedure [12]. Albumin proteins were extracted by deionized water with intermittent vortexing every 10 min for half an hour. The supernatant obtained following centrifugation at 2000 rpm for 5 min (albumin extract) was saved for further analysis. The procedure has been repeated twice to remove all the albumin fractions before proceeding to the next step.

The pellet from the final albumin extraction was used to extract globulin fraction. The extraction was conducted using 0.5 N NaCl for 30 min with intermittent vortexing every 10 min. In similar fashion, the supernatant obtained following centrifugation at 2000 rpm for 5 min (globulin extract) was kept in

Source: Jemanesh K. [13], Crop variety register Issue 1 - 12 and http://wheatatlas.org/ website CGIAR = Consultative Group for International Agricultural Research; CIMMYT = International Maize and Wheat Improvement Centre.

refrigerator for further analysis. Extraction was repeated twice to avoid carry over (cross-contamination) of globulin fraction to the subsequent steps. The centrifugate was washed with distilled water to reduce the effect of the salt in the extraction of other proteins in the subsequent steps.

Protein fractions were analyzed under reduced conditions (in the presence of beta-mercaptoethanol) using discontinuous SDS-PAGE [14]. The electrophoresis was conducted in 12% separating gel and and 6% stalking gel. Separation of protein bands was conducted at a constant voltage of 200 V until bromophenol blue passes the stacking gel and then raised gradually to 500 V. Following electrophoresis, the gels were stained with 0.2% (w/v) CBB in 45% (w/v) methanol and 10% (w/v) acetic acid under constant agitation. Destaining was conducted in 45% methanol and 10% acetic acid solution.

The acrylamide solution was made from 30 g acrylamide monomer and 0.8 g bisacrylamide in distilled water. The resolving gel buffer was composed of 36.6 g Tris base and 40 mL of 0.1 M HCl and the pH was adjusted to 8.8. Stacking gel buffer composition was 6.06 g Tris base dissolved in 40 mL of distilled water. The pH of the stalking gel adjusted to 6.8 using weak acid.

Molecular weights of the protein bands (polypeptides) were estimated by using Thermo Scientific PageRuler Plus prestained Protein Ladder having a mixture of nine blue, orange, and green dyed proteins (10 - 250 kDa).

The protein bands on the destained gel were quantitated using AlphaEaseFC 4.0 software (Alpha Innotech Corporation, San Leandro, CA).

The extracted samples were mixed with a buffer containing 2 mM DTT, 0.1% Triton X-100 and 63 mM Tris-HCl (pH 6.8). The reagent solution was prepared by mixing the Bradford reagent (100 mg of CBB G-250 in 50 ml of 95% ethanol) with 100 ml of 85% ortho-phosphoric on a magnetic stirrer. The resulting solution was filtered through filter paper (Whatman) and stored in a dark bottle at 4˚C. Standard curve was plotted by using seven concentrations [0 (blank), 0.2, 0.4, 0.6, 0.8, 1, 1.2 mg/ml] of bovine serum albumin (BSA) against a blank (deionized water). Absorbance was measured from 1 ml of the reaction solution at 595 nm after 3 min of incubation at room temperature. The absorbance of blank was subtracted from standards concentrations to obtain the actual concentration of the sample. Quantification of proteins was performed in triplicate. The spectrophotometric absorbance was read on Biotek Synergy 2 Micro-plate reader instrument using a Gen5 computer software program.

The absorbance reading was converted to protein concentration using a standard curve established with BSA dissolved in lysis buffer. The protein content was calculated using the following equation:

Protein ( % , w / w ) = C V D / M × 100

where C is protein concentration (mg/ml) obtained from standard curve.

V is volume (ml) of the lysis buffer used to resuspend the biomass.

D is the dilution factor.

M is the amount of biomass (mg).

The total protein content of the durum wheat varieties was estimated by micro-Kjeldahl method using nitrogen analyzer by taking 5.7 as a conversion factor. Three replicate measurements were taken to estimate error variance.

SDS-sedimentation volume was measured following Axford method [15].

All analyses were replicated three times and means were compared by Fisher’s least significant differences (LSD) at P < 0.05 and P < 0.01 using SAS statistical package [16]. Dendrogram was constructed from the electrophoretic data using Xlstat software.

The water soluble albumin and dilute salt soluble globulin proteins were separated and quantified. It is found that variety Robe had the highest albumin content (19.35%) while Boohai (15.34%) lowest. Albumin concentrations in wheat grain ranging from 18% to 21% have been reported elsewhere [17].

In globulins, the highest amount was recorded for varieties Cocorit-71 (10.70%), while for Gerardo (6.07%) the lowest. The albumin and globulin content of most common wheat proteins is reported to be in the range of 20% - 25% [2]. Our findings also confirmed similar results with an average albumin and globulin content of 27%.

The data presented in Table 2 showed that, the content of total protein is statistically significant (P < 0.01) suggesting the presence of considerable variation in protein content among durum wheat varieties. In the present study, protein content in grain of durum wheat ranged from 8.08% (Candate-Utuba) to 14.28% (Boohai and Tob-66). Other previous research findings reported grain protein content in the range of 7 to 12.5 among 15 Ethiopian durum wheat varieties [18]. The protein content of grain is affected mainly by genetic factor. However, environment and many other factors may also play a great role in determining the protein content of the crop [19] [20].

The experimental results based on the absorbance reading revealed gliadin proteins (also known as large monomeric gluten proteins), to be predominant protein fractions (51.47%). Albumins and globulins accounted only for 27% of the total protein. The ratio of gliadin to glutenin proteins was close to 4:1.

The correlation between total protein content and individual protein fractions was evaluated statistically (results not shown here). There was no significant correlation among the individual protein fractions and the total protein content. Contrary to the present finding, there are reports of strong relationship between albumin-globulin fractions and total protein content [21]. Sedimentation volume test had some positive correlation with protein content and this shows that the volume of sediment is a good indicator of protein content.

The albumin-globulin proteins were separated and identified by electrophoresis.

w/w, the proportion of individual protein fractions from total protein expressed in weight to weight basis.

The electrophorograms show the patterns and molecular weight distribution of albumin and globulin proteins (Figures 1-3).

Around 10 to 16 different polypeptides were detected in the globulin fraction. Their molecular weight ranged from 10 to 70 kDa. Among the seven prominentglobulin polypeptides identified, five major ones’ molecular weight were 15, 20, 35, 55, and 70 kDa, respectively (Figure 1). These results are consistent with previous research findings [3] [22] which reported globulin polypeptides ranging from 12.4 to 76.4 kDa.

In the globulin proteins, high polymorphism in both band intensity and occurrence was observed in those with molecular weights between 15 and 35 kDa. The presence of polymorphism in this molecular weight range suggests that

globulins could also be used as suitable and useful genetic markers to discriminate genotypes.

A major cause of polymorphismin any specific protein is known to be related with gene silencing [23].

The gel electrophoresis result of albumin proteins did not show any significant variation among the genotypes. Almost all bands were monomorphic across all wheat genotypes. Molecular weight of the polypeptides ranged from 10 to 65 kDa and the number and position of bands were similar for all varieties (Figure 2). Generally, around 20 different polypeptides were detected in the albumin fraction. Likewise, the number of bands in globulin proteins was from 10 to 16. This is in agreement with other previous research findings [3].

The heterogeneity of albumin and globulin was also revealed by the banding patterns of the mixture of globulin and albumin fractions (Figure 3). The varietal heterogeneity of these fractions was more similar to that detected by globulin fractions. Diversity of albumin and globulin proteins in wheat varieties was demonstrated in the earlier research works [24]. This pioneering research on albumins has reported the presence of three protein components in the water soluble proteins. Following this finding, there were some contradicting reports on the presence of noted molecular weight heterogeneity in both in the water-soluble and salt-soluble proteins.

The dendrogram constructed based on the electrophoretic data of globulin protein fractions of the 20 durum wheat varieties is shown in Figure 4. There was fairly clear separation of the varieties. The cluster analysis result of globulin proteins classified the durum wheat varieties into three major lineage groups at a cutoff value 2.5. The varieties in cluster 3 were composed of widely adapted types, whereas the varieties in cluster 1 and cluster 2 were primarily from highland areas.