Durum Wheat Bread with a Potentially High Health Value through the Addition of Durum Wheat Thin Bran or Barley Flour

The enrichment of semolina bread with prebiotic ingredients such as β-glucans may exert health-promoting effects. This work presents the results of a general recipe development aimed at improving the nutritional value of bakery products. In this study, increasing amounts (0%, 2%, 5%, 7%, and 10%) of thin bran or barley flour were added into re-milled durum wheat semolina to prepare breads. The technological quality of doughs and breads was investigated. In general, the Farinograph water absorption of flour and dough stability increased with increasing inclusion levels of barley flour or thin bran (up to 73.23% and 18.75 min, respectively), contrarily to the increase of dough development time only in barley inclusion (4.55 min). At the same time, the softening index decreased for almost all of these, except for 2% of thin bran or barley flour inclusion. At Mixograph, mixing time increased (up to 5.13 min) whilst the peak height decreased. The specific volume and hardness of loaf differently decreased for almost all thesis (ranges 12.6–24.0% and 39.4–45.5%, respectively). The other quality parameters remained unchanged compared with semolina bread. After baking, β-glucan levels increased differently at all the inclusion levels (2.35-fold, on average). The breadcrumb color was deep brown, while the crust became lighter in color. The breads contain β-glucans even at low percentages of barley/bran inclusions while maintaining their technological performance. In conclusion, the results show an interesting potential of barley flour or thin bran as ingredients in breadmaking to increase the β-glucans daily intake, but further investigations are needed to achieve improved quality features.


Introduction
Semolina is a typical cereal product of the Mediterranean region derived from the endosperm of the grain, presenting low levels of some valuable, healthy compounds, such as dietary fiber, vitamins, minerals, and antioxidants, which are abundant in bran [1]. Enriching wheat-based products, like bread, with health-promoting compounds, such as prebiotics, is becoming a common approach for the development of functional foods [2].

β-Glucans and Chemical Characterization of Semolina and Flour Blends
The incorporation of β-glucan in foods such as bread and pasta, seen as highly healthy by the consumer [21], through the use of barley flour [22] in progressive replacement of wheat flour [23] has been widely studied, as regards the ability to interfere on the physicochemical and rheological properties of doughs.
The analyses conducted in our study showed high β-glucan contents in 100% barley flour, in line with what was found by other authors [24][25][26], equal to 10.61% (Table 1). As regards the content of β-glucan in semolina (0.31%), it has clearly lower values than 100% barley flour, tending to increase in content in relation to the presence of the bran fraction present as observed by other authors [27], confirming the values we reported for 100% thin bran (1.19%). Naturally, the addition of barley flour in the various supplements resulted in an increase in the content, measured in the 10% supplemented semolina (1.58%), in β-glucan equal to five times compared to the 100% control re-milled semolina.
This differs from what was found in thin bran, which did not contribute significant quantities of β-glucan in the various additions. These data agreed with Basman et al. [28], who, studying the effect of barley flour and wheat bran supplementation on the composition of Turkish flatbread, observed only a significant increase in β-glucan values with an increasing percentage of barley flour.
Moisture, protein content, and ash of semolina fulfilled the legal requirements (Italian Presidential Decree n. 187/2001) and were in the range observed by [15,29]. Moisture did not show differences among barley flour or thin bran inclusions, whilst resulting lowest in barley flour and the highest in thin bran whole flours. Protein content decreased significantly in whole barley flour and different barley blends at increasing percentages of inclusion compared with semolina, accordingly to Mohebbi et al. [30]; on the contrary, in thin bran, protein content was higher than semolina, at increasing levels of inclusion. Otherwise, in wheat flour complemented with different amounts of oat flour (5%, 15%, and 25%), the protein content was similar with up to 15% oat inclusion [31]. These data confirmed that a relatively greater proportion of proteins among cereals was found in wheat flours (10.18-11.25%) compared with barley flours (7.09-9.04%) [29,32].
Similar to durum wheat thin bran, barley flour is also a source of ash. The aleurone cells, together with the testa and germs, contain essential minerals required for embryo growth [33]. The highest ash values were observed for whole thin bran and whole barley flours. The mean ash content increased with an increase in barley or thin bran inclusions, remaining below the legal limits of Italian law for semolina (maximum ash content of 0.9%). The data observed were in line with those reported by Mehfooz et al. [34] for wheat flour/barley blends.
Regarding the color indices (Table 1, Figure 1), significant differences in the brown index were observed for whole thin bran flour and 10% thin bran inclusion compared with semolina, whilst the red index was less negative in whole barley flour and highly positive in whole thin bran flour; finally, significant differences in the yellow index were observed only between whole barley and whole thin bran flours. These data were in accordance with Basman et al. [28]. observed only between whole barley and whole thin bran flours. These data were in accordance with Basman et al. [28].  Table 2 shows a two-factor ANOVA (analysis of variance) of the physicochemical features of the thin bran and barley flour. Almost all variables, except β-glucan content, which is about six times higher, show higher values in thin bran. Table 3 shows a two-factor ANOVA (analysis of variance) of the physicochemical features to the different percentages of integration of two flours.  [35], who observed an increase in the mixing time and a decrease in protein content because of the addition of non-wheat protein fractions. Instead, in thin bran inclusions, higher mixing time values were observed for 2% and 10% and lower values for 5% and 7% as compared with semolina. Different mixer inclusions probably influence the mixing procedures and dough properties differently. Peak dough height (i.e., a measure of dough strength) was lower in barley and thin bran formulations compared with semolina, with different behaviors except for 2% barley and 10% thin bran, which were similar to semolina. In particular, barley at the major percentage of inclusion corresponded lower peak dough height  Table 2 shows a two-factor ANOVA (analysis of variance) of the physicochemical features of the thin bran and barley flour. Almost all variables, except β-glucan content, which is about six times higher, show higher values in thin bran. Table 3 shows a two-factor ANOVA (analysis of variance) of the physicochemical features to the different percentages of integration of two flours. β-glucan increases with the increasing percentage of supplementation until it reaches 5.9% in 100% supplementation.
The technological properties of dough from barley-or thin bran-semolina flour blends compared with semolina were determined by Mixograph and Farinograph tests (Table 4). Mixing time (namely, the time in minutes required for optimum dough development) increased in barley with the increasing percentage of inclusion. The results were in line with those reported by Tömösközi et al. [35], who observed an increase in the mixing time and a decrease in protein content because of the addition of non-wheat protein fractions. Instead, in thin bran inclusions, higher mixing time values were observed for 2% and 10% and lower values for 5% and 7% as compared with semolina. Different mixer inclusions probably influence the mixing procedures and dough properties differently. Peak dough height (i.e., a measure of dough strength) was lower in barley and thin bran formulations compared with semolina, with different behaviors except for 2% barley and 10% thin bran, which were similar to semolina. In particular, barley at the major percentage of inclusion corresponded lower peak dough height compared with semolina; the contrary was observed for thin bran. The high height values can probably be attributed to the influence of flour protein content on flour water absorption, so with increasing protein content, doughs became stiffer, resulting in increasing Mixograph peak height values. The opposite was for dough at low protein content. A positive relationship between protein content and Mixograph peak height (r = 0.60 **) was observed, and it was consistent with previous studies [36]. Mixograph peak height may provide an important quality criterion in assessing flour quality performance, especially in early-generation selection. Regarding the Farinograph parameters, water absorption (i.e., the percentage of water required to reach a dough consistency of 500 Brabender Units) progressively increased as the amount of barley and thin bran integration added increased, and barley inclusion caused a greater increase (from 63.05% to 73.23% in barley and from 63.68% to 65.32% in thin bran, respectively). The increase of hydration capacity was different between flour inclusions, which could be explained by the intake and type of fibers that barley or bran brought into the dough, also determining higher yield in breads [37]. This result has been related to the high-water absorbing capacity of the dietary fiber and its ability to compete for water with other components in the dough system, interfering with the formation of a strong gluten network and dough stability [26]. Similar to Kaur et al. [36], a strong relationship (r = 0.80 **) between Mixograph peak height and Farinograph absorption was evidenced, which could be due to the impact of proteins as well as other factors such as starch damage and gluten strength which can strongly influence Farinograph absorption.
Dough development time (i.e., the time needed from the first addition of water to reach the maximum consistency) did not have a clear-cut trend with the increase of barley or bran flour supplementation. In addition, bran integrations did not show statistically significant differences compared with semolina (p < 0.01), except for 2% bran inclusion, showing the lowest development time. The results were in line with Popa et al.'s findings [37]. The development time of different kinds of flours could be strongly influenced by protein content and quality, as evidenced by other authors [30,38].
Dough stability (i.e., the difference between the time needed to reach the dough consistency of 500 Brabender Units and the time when it leaves the 500 B.U. line) was maximum at 5% added of barley flour (10.08 min), then tended to decrease at larger additions of barley flour. Instead, it was high and constant from 5-10% bran inclusion (18.75 min, on average), reaching equal value to the control at 2% level (3.78 min). The different effects in dough stability among samples could depend on the development of stiffer dough with the increase in barley (reaching a plateau at 5% inclusion) or bran (at 5-10% inclusions) due to more amount of water being absorbed by fiber [34]. Since stability is correlated with dough tolerance to kneading, fermentation, and even with the volume of finished products, the evolution of this parameter can be used to determine the optimal amount of flour inclusion. Finally, the softening degree (i.e., the loss of dough consistency after 12 min) significantly decreased in all samples compared with semolina, except for 2% levels in barley and thin bran. The reduction of softening index could be due to a dilution of gluten by bran leading to dough deterioration, breaking of the starch-gluten network structure determining a decrease of consistency, with the release of water from the system [39]. On the other hand, the discrepancies related to the influence of fiber on the dough may arise from differences in the molecular size, solubility, and concentration range of the polysaccharides, as well as the flour types used for supplementation [40]. Table 5 shows a two-factor ANOVA (analysis of variance) of the main technological parameters of the thin bran and barley flour. All variables, except dough stability, show higher values in barley flour.  Table 6 shows a two-factor ANOVA (analysis of variance) referring to the different integration percentages of two flours: thin bran and barley flour. As far as the Mixograph is concerned, as the percentage of integration increases, the mixing time increases, and the peak dough height decreases. The Farinograph parameters increase as the integration percentage increases, except for development time and dough stability at a 10% inclusion level.

The Quality Parameters of Breads Using Different Formulations
Significant differences (p < 0.01) in specific volume were shown among the bread samples, with slight changes across the type and level of supplementation (Table 7, Figure 2). A clear difference, in terms of a decrease equal to 24% on average, was observed at 2% and 10% barley inclusions as well as at 10% bran inclusion as compared with semolina bread. Considering the other percentage of inclusions, an average decrease in specific volumes of 15.5% and 12.6% for barley and bran, respectively, compared with semolina bread, was evidenced. The results of this study were in line with other authors [41][42][43], who observed slight decreases or similar values to the control sample of specific volume in wheat breads incorporating barley/soybean flours or chia seeds. The differences observed among samples could be due to the introduction of fiber-rich products into the dough that negatively affects the formation of gluten, reducing its ability to retain gases [44][45][46]. Instead, others [40,47] found higher bread volume due to the high MW β-glucans found in barley/oat flours or added as β-glucan isolate, which may, in turn, stabilize gas cells by increasing the viscosity of the doughs. Table 7. Evaluation of physical properties and moisture of the bread samples produced using different types and levels of supplementation: one-factor ANOVA (analysis of variance) (data are means ± standard deviations). in wheat breads incorporating barley/soybean flours or chia seeds. The differences observed among samples could be due to the introduction of fiber-rich products into the dough that negatively affects the formation of gluten, reducing its ability to retain gases [44][45][46]. Instead, others [40,47] found higher bread volume due to the high MW β-glucans found in barley/oat flours or added as β-glucan isolate, which may, in turn, stabilize gas cells by increasing the viscosity of the doughs.  The same behavior observed for specific volumes was also found for height and specific weight among the control and the other bread samples.

Sample
No significant differences in crumb porosity were observed among the bread samples compared with semolina bread, except for the bread containing 2% thin bran supplementation, in which the crumb porosity suffers deterioration, showing a nonhomogeneous crumb. The results of this study disagree with [48,49], who observed increases in the porosity of the crumb, measured on the Dallmann scale, replacing wheat flour with brewer's spent grain or fresh pumpkin pulp.
A decrease in hardness was observed in breads with barley/thin bran inclusion compared with bread control. Barley or thin bran addition to semolina decreased the The same behavior observed for specific volumes was also found for height and specific weight among the control and the other bread samples.
No significant differences in crumb porosity were observed among the bread samples compared with semolina bread, except for the bread containing 2% thin bran supplementation, in which the crumb porosity suffers deterioration, showing a non-homogeneous crumb. The results of this study disagree with [48,49], who observed increases in the porosity of the crumb, measured on the Dallmann scale, replacing wheat flour with brewer's spent grain or fresh pumpkin pulp.
A decrease in hardness was observed in breads with barley/thin bran inclusion compared with bread control. Barley or thin bran addition to semolina decreased the hardness by 45.5%, on average at 5% and 7% barley inclusion, and by 39.4%, on average at 2% and 5% thin bran inclusion, compared with semolina bread. The results of this study agreed with Adamczyk et al. [41], who observed a reduction of hardness in bread by replacing 1 or 5% w/w whole chia seeds with wheat flour.
Finally, the moisture content ranged from 19.32-28.33% in barley supplementations, increasing with the increase of barley amount added up to reach similar values to bread control at 10% of barley inclusion (28.33% vs. 28.47%). In contrast, the inclusion of increasing thin bran levels determined a decrease in moisture, showing absolute values higher than semolina bread (30.35% vs. 28.47%, on average). Presumably, the different water hydration properties and profiles of the fiber blend polymers might justify the different behavior in water retention capacity and moisture of breads, reducing bread hardness [50,51]. Table 8 shows the evaluation of the physical properties of the bread samples produced using different types of supplementation: barley flour and thin bran. Two-factor ANOVA was not significant. Table 8. Evaluation of physical properties of the bread samples produced using different types of supplementation: two-factor ANOVA (analysis of variance) referred to the two flours: barley flour and thin bran (data are means ± standard deviations).  Table 9 shows the evaluation of the physical properties of the bread samples produced using different levels of supplementation: two-factor ANOVA (analysis of variance) referred to the different percentages of integration of two flours. The specific volume and height of the loaves decrease as the percentage of integration increases. The specific weight and moisture parameters show the opposite trend. The hardness remained almost unchanged, while the porosity was not significant. Table 9. Evaluation of physical properties of the bread samples produced using different levels of supplementation: two-factor ANOVA (analysis of variance) referred to the different percentages of integration of two flours (data are means ± standard deviations).

Results of β-Glucans of Breads with Different Percent of Barley/Thin Bran Flour Compared with Control Bread
Concerning the β-glucans content, after baking, supplemented bread showed an increase in β-glucans compared with semolina bread (Table 10), independently from the enrichment level. In fact, the bread obtained with the inclusion of barley or thin bran showed twofold β-glucan increases at 2% barley supplementation up to threefold at 10% barley, while 1.7-fold and 2.7-fold increases were recorded at 7% and 10% thin bran supplementation. Indeed, it is well documented that many processing methods, such as milling, germination, cooking, baking, extrusion roasting, and freezing, can affect the stability, solubility, and viscosity of β-glucan differently [52]. Johansson et al. [53] observed decreases of β-glucan after baking, whilst Cavallero et al. [54] and Blandino et al. [55] reported that the bread-baking process did not reduce the β-glucan content. The increased values reported in this study, as observed by Izydorczyk et al. [56], could be due to hydrothermal treatment (steaming) during baking that, although not affecting the extractability of β-glucans, prevent their enzymatic hydrolysis, with no change or disrupt the β-glucan to form other aggregates. The same result was observed by other authors [57] in pasta obtained by replacing semolina with barley flour rich in β-glucan, in which cooking increased the extractability and the viscosity, determining its physiological effectiveness.  Table 11 shows the β-glucans contents (% w/w) of the enriched breads, based on the raw materials (thin bran and barley flour), as determined by the two-factor ANOVA. ANOVA was not significant. Table 11. β-Glucans contents (% w/w) of the enriched breads, based on the raw materials (thin bran and barley flour), as determined by the two-factor ANOVA (analysis of variance).  Table 12 shows the β-glucans contents (% w/w) of the enriched breads, based on the raw materials (thin bran and barley flour) and on the level of inclusions, as determined by the two-factor ANOVA. As expected, as the percentage of integration increases, the content in β-glucans increases. Table 12. β-Glucans contents (% w/w) of the enriched breads, based on the raw materials (thin bran and barley flour) based on the level of inclusions, as determined by the two-factor ANOVA (analysis of variance).

Color Indices in Crumb and Crust Breads Obtained with Different Formulations
Other important features of the bread samples are the crust and crumb color, which are also highly associated with bread consumers' acceptance. Crumb color results (Table 13) reveal that the inclusion of barley or thin bran increased the brownness tone (100-L* values), the reddish tone (a* values), and the yellowness tone (b* values) in different ways. In contrast, crust color lighter breads (100-L* values) with a remarkable yellowness tone (b* values) and reddish tone (a* values) were obtained. The brownness tone in the crumb was highest at 10% barley inclusion, while the color at 2% barley level was similar to semolina bread; the brownness tone in the crust of bread with 10% barley inclusion was similar to semolina and lower in the other samples. The brownness, which is influenced by the flour type and extraction [58], was the result of the occurrence of a Maillard reaction during heat treatment and of the enzymatic oxidation-polyphenol oxidase and peroxidase-of phenolics to brown quinones [59]. The reddish tone in the crumb varied from less negative to strictly positive and/or higher values than the semolina bread ones, with a maximum obtained with 10% barley inclusion. Instead, the crust's maximum value was reached at 2% bran inclusion. The yellowness tone increased differently at different flour inclusion, both in the crumb and crust.  Table 14 shows the evaluation of colorimetric parameters of the bread samples based on different types of supplementation, as determined by the two-factor ANOVA (analysis of variance). ANOVA was not significant.  Table 15 shows the evaluation of colorimetric parameters of the bread samples based on the level of inclusions, as determined by the two-factor ANOVA (analysis of variance). The three color parameters of the crumb increase as the percentage of integration increases. Regarding the crust, the brown index follows the same trend, while the red and yellow indexes decrease as the percentage of integration increases.

Physico-Chemical Analyses of Raw Materials and Flour Blends
The β-glucan content of semolina and flour blends was determined enzymatically according to AACC 32.23.01 method [60] using the Megazyme β-glucan assay kit (Megazyme, Bray, Ireland) and was expressed as the percentage of flour weight on fresh weight (f.w.) basis.
Protein content was determined using the Kjeldahl method, according to the American Association of Cereal Chemists (AACC) approved method 46-13.01 [61]. The multiplication factors used were 5.7 for cereals.
Ash content was obtained following the ISO method 2171 [62]. The color parameters in the color space L*, a*, and b* were determined by Chroma meter CR-300 (Minolta, Osaka, Japan) under the illuminant D 65 . Brown index was calculated as 100-L*.
The analyses of the raw materials and flour blends were carried out in triplicate.

Technological Tests on Doughs of Semolina and Flour Blends
The Mixograph curves were obtained following the AACC method 54 The measures were replicated three times.

Baking Test
The breadmaking test was performed on the semolina control and on flours obtained from each thesis according to the AACC 10-10.03 procedure (2000), as modified for durum wheat by Boggini and Pogna [65], into the baking time (18 min) and temperature (217 ± 4.24 • C), to obtain two loaves of about 140 g/each. Two independent replicate baking experiments were carried out.
Hence, a total of 36 loaves were obtained, onto which the following traits were individually measured: volume, height, weight, crumb porosity, hardness, moisture, crumb, and crust color.
The specific volume and specific weight were calculated by comparing the loaf volume to its weight and the loaf weight to its volume. The loaf height was measured using a digital caliper (Digi-MaxTM, SciencewareR, NJ, USA). The crumb porosity was assessed according to the Dallmann scale [66].
The loaf hardness was measured using a texture analyzer (Zwick Z 0.5 Roell, Ulm, Germany) equipped with an aluminum 8 mm diameter cylindrical probe.
The moisture content was determined by gravimetric analysis. The CIE L*, a*, b* color parameters were measured for the crumbs in the transversely cut bread and on the crust surface, using a Chroma Meter (CR-200, Minolta) with illuminant D 65 . The measurements were replicated twice, except for the loaf hardness, which had in triplicate. The results of two loaves of a single batch for each thesis were averaged into one replicate value.

Statistical Analysis of Data
The statistical analysis was performed using the Statgraphics ® Centurion XVI software package (Statpoint Technologies, INC., The Plains, Virginia). One-factor and two-factor analysis of variance (ANOVA), followed by Tukey's HSD test (p ≤ 0.05), was carried out on all physicochemical, technological, and breadmaking attributes. The two factors were considered: 1. the type of ingredient, 2. the amount of ingredient. A one-factor analysis determined the interaction of the factors studied, while a two-factor analysis analyzed each factor's influence (or lack of influence) individually.

Conclusions
In general, although the Mixograph and Farinograph parameters were negatively affected by the inclusion of barley flour or thin bran and even exceeded the typical values of semolina control [56], the final bread quality showed a little reduction in specific volume, which is an important parameter for evaluating bread-making quality. Also, a decrease in hardness was observed, while other parameters, such as crumb porosity, remained unchanged.
After baking, an increase of β-glucans was observed in all samples, more evident at high barley flour/thin bran inclusion levels, showing that the heat treatment inactivated endogenous enzymes resulting in reduced β-glucans degradation.
Moreover, the variation of color tones, already significant at low levels of supplementation, was generally progressively more evident with the increase of flour added. As color is an important attribute that strongly influences consumer choice, high differences from bread without supplementation could be negatively considered. So, the inclusion of intermediate percentages of alternative cereals to semolina formulations could be a compromise to improve the nutritional properties of breads while maintaining the rheological performance and acceptability in terms of color similarity to semolina bread.

Future Work
The study conducted so far has to be intended as a work to establish the technological conditions for optimal final product development. This is to identify the type of flour to add, different from re-milled semolina, giving the best results and the optimal percentage to be used. In the future, we will focus on bread's nutritional, nutraceutical, sensory, and storage aspects, particularly the amount of extracted β-glucan and the glycemic responses to carbohydrate products.