ABSTRACT: The formulation of gluten-free bread, which will be suitable for patients with coeliac disease, was optimized to provide bread similar to French bread. The effects of the presence of hydrocolloids and the substitution of the flour basis by flour or proteins from different sources were studied. The added ingredients were (1) hydrocolloids (carboxymethylcellulose [CMC], guar gum, hydroxypropylmethylcellulose [HPMC], and xanthan gum), and (2) substitutes (buckwheat flour, whole egg powder, and whey proteins). The bread quality parameters measured were specific volume, dry matter of bread, crust color, crumb hardness, and gas cell size distribution. Specific volume was increased by guar gum and HPMC. Breads with guar gum had color characteristics similar to French bread. Hardness decreased with the addition of hydrocolloids, especially HPMC and guar. Breads with guar gum had the most heterogeneous cell size distribution, and guar gum was therefore selected for further formulations. Bread prepared with buckwheat flour had improved quality: an increased specific volume, a softer texture, color characteristics, and gas-cell size distribution similar to French bread. Bread with 1.9% guar gum (w/w, total flour basis) and 5% buckwheat flour (of all flours and substitutes) mimicked French bread quality attributes.
Coeliac disease (CD) is a pathology affecting the upper small intestine mucosa due to an inappropriate immune response to gluten protein fractions (Marsh 1992), which are mainly present in wheat, barley, and rye. Patients with CD suffer from symptoms such as diarrhea, weight loss, and iron, folate, or vitamins B12 and D deficiencies (Woodward 2007). In Europe and the United States, the prevalence of CD is estimated at 1 for 100 to 200 inhabitants (Cook and others 2000; Hill and others 2000). Despite considerable scientific progress in understanding CD and in preventing or curing its manifestations, to date a strict gluten-free diet for life is the only treatment for CD patients (Niewinsky 2008).
In many countries, bread is mainly prepared from wheat flour containing proteins implicated in CD, not only gliadin peptides but also glutenin peptides (Howdle 2006). These proteins play a key role in the unique baking quality of wheat by conferring an appropriate water absorption capacity, cohesiveness, viscosity, and elasticity to the dough (Wieser 2007). During the fermentation step, part of the carbon dioxide, produced by yeast, is diffused to the micro alveoli (incorporated at the kneading step), which become gas cells. At advanced stages of fermentation, these gas cells are separated by a fragile liquid lamella stretching across the holes in a starch–gluten matrix, which contributes to the development of a porous structure in breads (Gan and others 1995).
There is still a need to find substances that could improve the quality of gluten-free breads. Gluten-free bread formulations with starches (Sanchez and others 2002), hydrocolloids (Kulp and others 1974; Cato and others 2004; Lazaridou and others 2007; Anton and Artfield 2008), and/or protein sources (Moore and others 2004; Ahlborn and others 2005) have been investigated for more than 40 y. Among the protein sources used in these formulations for their technological benefits, buckwheat flour also has the nutritional advantages of containing dietary fiber, resistant starch, and protein of high nutritional value due to its relatively high lysine content (Wijngaard and Arendt 2006). The main objective of these studies was to determine the effect of the type and quantity of supplement on improving the quality of gluten-free breads. However, the bread reference containing gluten can vary from one country to another. In France, there is still a strong attachment to French bread. Beyond legislative aspects, which identify the ingredients and process required to call bread a “French bread,” this type of bread is characterized by unique properties such as a crusty and well-colored crust, and a cream crumb with gas cells of heterogeneous size (Roussel and Chiron 2005). The main objective of our study was to develop an optimized gluten-free bread formulation to obtain a bread similar to French-style breads, and also to study the impact of hydrocolloids and protein sources on bread quality. Therefore, 2 sets of trials were conducted to reduce the differences between such products. First, the effects of hydrocolloids on bread quality were investigated. Four hydrocolloids were tested: carboxymethylcellulose (CMC), guar gum, hydroxypropylmethylcellulose (HPMC), and xanthan gum. From this stage, guar gum, which provided breads closest in quality to French-style breads, was selected for the following set of trials. In the 2nd stage, different protein sources (buckwheat flour, whole egg powder, and whey proteins) were introduced and their impact on bread quality was evaluated.
Materials and Methods
Rice flour, cornflour, potato starch, cornstarch, compressed yeast, salt, sunflower oil, and tap water were used in all bread formulations. Rice flour and cornflour were from Livrac (Haute-Goulaine, France). Potato starch was from KMC (Brande, Denmark), and cornstarch from Tate and Lyle (Aaist, Belgium). Compressed yeast, salt, and sunflower oil were obtained from local commercial sources (Nantes, France). CMC (Akzo Nobel, Amersfoort, The Netherlands), HPMC (DOW Europe GmbH, Stade, Germany), guar gum (Danisco, Zaandam, The Netherlands), and xanthan gum (Cargill, Baupte, France) were tested. Buckwheat flour (Bio Moulin, Boussay, France), whey proteins (Eurial, Nantes, France), and whole egg powder (Ovifrance, Plaintel, France) were studied. The formulations used in the preparation of breads are shown in Table 1.
Table 1—. Gluten-free bread recipes.
The formulation without any hydrocolloid or protein source was named Control. The other formulations were named after the added hydrocolloid or protein source, for instance XG corresponded to the formulation containing xanthan gum and EP to that with egg powder. Flours (including the protein source when added), starches, hydrocolloid (if added), and salt were blended at speed 1 (46 rpm) for 10 s in a mixer (Kitchenaid, St. Joseph, Mich., U.S.A.). Compressed yeast was incorporated and the ingredients were blended again at speed 1 for 10 s. Sunflower oil and tap water (at 20 °C) were added and all ingredients were mixed at speed 2 (82 rpm) for 2 min. Dough pieces (70 g) were weighed and placed in individual aluminum muffin-like pans (35 mm height; 75-mm top diameter; 68-mm bottom diameter) lined with baking paper and were proofed for 50 min in a fermentation cabinet (Hengel, Le Coteau, France) at 40 °C and 95% humidity. Just before baking, the dough surface was cut twice with a razor blade then breads were baked (oven condo Miwe, Arnstein, Germany) at 200 °C for 40 min with 0.5 L of steam at the start of baking. Breads were cooled at room temperature for 1 h and then placed in sealed plastic bags at room temperature for 1 h before analysis. An example of the final product BF is shown in Figure 1. Breadmaking was duplicated for each formulation tested.
Fresh French (FR) bread purchased at a local bakery (Nantes, France) was used as the French bread reference and was analyzed in the same way as gluten-free breads.
For dry matter determination, 3 ± 0.3 g of crushed bread was weighed into aluminum dishes and dried for 24 h in an oven (Memmert, Schwabach, Germany) at 103 °C. Five replicates were carried out. Crumb characteristics were assessed using a texturometer (Lloyd Instruments LR5K, Southampton, UK). A compression test with a cylindrical probe (20 mm diameter) was performed. 15 mm-thick slices were cut from the middle of the bread using an electric slicer (Graef, Arnsberg, Germany). Crumb was compressed to 40% of its initial height at 5 mm/s. Tests were performed on 2 slices from each of 4 breads. Crust color was measured on the top surface of 4 breads using a Minolta chromameter (Minolta CR 400, Osaka, Japan). L*, C*, and h were recorded. Specific volume was measured on 3 breads using rapeseed displacement. Images of sliced breads were captured using a flatbed scanner (HP, Palo Alto, Calif., U.S.A.). These images (10 × 10 cm) were scanned at 350 dpi and analyzed at gray levels using Ccell (Campden & Chorleywood Food Research Assn., Gloucestershire, UK) (Figure 1). The number of cells per square centimeter and the porosity (cell area × 100/slice area) were recorded for each slice. Along with the gas cell size distribution, the area and mean diameter of each gas cell (shape was assimilated as a perfect circle) were recorded. Eleven categories of gas cells (numbered from 1 to 11) were determined according to their diameter (from 0 to above 10, with an amplitude of 1 mm per category: diameter < 1 mm, 1 mm ≤ diameter < 2 mm , … , diameter ≥ 10 mm) and for each category, the percentage of gas cell area was calculated. Four slices of bread were analyzed for each breadmaking.
Data concerning the 2 sets of trials (effects of hydrocolloids and effects of protein sources) were analyzed separately. The statistical analysis for the 1st stage included Control, CMC, GG, HPMC, XG, and FR while for the 2nd stage it included GG, BF, EP, WP, and FR. These 2 datasets were analyzed using one-way analysis of variance (ANOVA) and the means were compared by the Fisher LSD test at a significance level of 0.05 in both cases with Statgraphics Plus 5.1 Software (Statistical Graphics Corp., Princeton, N.J., U.S.A.).
Results and Discussion
Gluten-free and French breads were evaluated for their specific volume and bread dry matter, crust color, crumb hardness, and gas cell distribution. French bread characteristics were considered as the targets to aim for with gluten-free breads.
Data concerning the 2 sets of trials (effects of hydrocolloids and effects of protein sources) are shown in the same figures or tables, although they are analyzed and discussed separately. Lowercase letters (a, b, c , …) are used to identify statistically different groups between the 1st set of trials (Control, CMC, GG, HPMC, XG, and FR) and uppercase letters (A, B, C , …) between the 2nd set of trials (GG, BF, EP, WP, and FR). At the end of the 1st set of trials, the GG formulation was selected for the 2nd set. GG was thus considered as a new basic formulation to be improved by the protein sources and was used to quantify the effect of each protein source.
Effects of hydrocolloids
Based on previous publications (Lopez and others 2004; Moore and others 2004; McCarthy and others 2005; Schober and others 2005; Sciarini and others 2007), preliminary baking trials were performed to determine the level of each hydrocolloid necessary to obtain dough that could form small baguettes (not too liquid and not too thick). The formulation without any hydrocolloid, called the Control, was developed during preliminary tests with the ingredients and level of supplement usually used in gluten-free bread formulations. It produced a liquid batter. To allow comparison with the Control and to get a precise idea about the effect of hydrocolloids on bread characteristics, all breads obtained from the other formulations were made in muffin-like pans.
Hydrocolloid addition had a significant impact on bread specific volume and crumb hardness when compared to the Control and French breads (Figure 2 and 3).
A specific volume of 2.47 ± 0.11 mL/g and 5.28 ± 0.30 mL/g was obtained for the Control and French bread, respectively, illustrating that the French bread was more than 2-times more voluminous as the Control bread. GG and HPMC breads had an improved specific volume of 2.82 and 3.33 mL/g, respectively, compared to the Control bread. Addition of CMC and XG did not significantly improve the specific volume of the breads compared to the Control bread. Lazaridou and others (2007) also reported that the incorporation of xanthan gum at 1% into gluten-free breads based on rice flour, cornstarch, and sodium caseinate did not change the specific volume; it was even decreased when xanthan gum was added at 2%. Thus, addition of this hydrocolloid makes the dough system too rigid to incorporate gases and produce voluminous breads. Hardness is the force at the compression peak. The Control breadcrumb, with a hardness of 6.01 ± 0.63 N, was considerably harder than French-style bread (0.61 ± 0.21 N). Hydrocolloid addition to the gluten-free formulation led to the production of softer crumbs, depending on the hydrocolloid type. The softest crumbs were obtained with HPMC and GG, 1.86 and 2.91 N, respectively. Softness was significantly and positively correlated with specific volume at 79%; voluminous breads (HPMC and GG) were softer than compact breads.
The number of cells per square centimeter increased by a factor of 2 between FR and the Control crumbs (Figure 4). The addition of CMC and XG did not modify this value compared to the Control. HPMC increased the number of cells in the crumb compared to the Control, while GG decreased it, suggesting that only guar gum addition gave a similar number of cells per square centimeter to French bread.
Porosity was obtained by dividing the total cell area by the slice area. It was higher for French-style bread than for the Control, 47.73%± 0.92% and 41.40%± 0.24%, respectively (Figure 5), illustrating that the area occupied by the cells was higher in FR crumb than in Control crumb, although they were less numerous. Addition of HPMC to the gluten-free formulation led to a decrease in porosity. Both a lower porosity and a higher number of cells than for the Control were obtained with HPMC. This result illustrates that there were more cells for a smaller area indicating that HPMC addition gave a crumb with a foamy structure. For this hydrocolloid, the porosity result contradicts the specific volume result. Indeed, HPMC addition increased the specific volume but decreased the porosity. One explanation could be that, due to the foamy structure of the crumb, the contrast is slightly lower between the gas cells and cell walls in this type of foamy bread; the gas cell area could be underestimated when compared to the other bread structures and thus porosity could be more difficult to evaluate by this method. Breads with CMC had a higher porosity than the Control and the closest to FR, and a number of cells similar to the Control; thus gas cells were bigger with this hydrocolloid and its addition could contribute to improving gas retention in cells. Guar gum and xanthan gum had a very slight impact on porosity. While XG bread had the same number of cells as the Control bread, GG bread had fewer cells. This result suggests that gas cells were bigger with guar gum addition than in the Control. Guar gum could modify dough viscosity in the early stages and then the GG dough would incorporate fewer cells than that of the Control. In this case, despite the low gas cell incorporation at the early stage, guar gum would have the capacity to improve gas retention at the fermentation step. One other explanation could be that, with the same capacity of gas cell incorporation at the early stages for GG and the Control, gas cells could coalesce with guar gum addition. CMC and guar gum were therefore the most efficient hydrocolloids in retaining gas in the bread.
Concerning the size distribution of gas cells, only 3 categories among the 11 showed significant differences between the formulations tested (Table 2); gas cells with a diameter below 1 mm (category nr 1), from 5 to 6 mm (category nr 6), and above 10 mm (category nr 11).
Table 2—. Cell size distribution (in cell area relative percentages) of gluten-free breads (Control, CMC, GG, HPMC, XG, BF, EP, and WP) and French bread (FR). Superscript lowercase letters and superscript uppercase letters were used to identify significant differences in the 1st and 2nd set of trials, respectively.
The relative area occupied by category nr 1 cells was higher for the Control than for French bread, 29.7%± 3.0% and 20.3%± 0.7%, respectively. Category nr 1 of HPMC breads was similar to the Control while for CMC breads it was similar to FR bread. XG and GG breads had an intermediate position with a relative area occupied by category nr 1 cells ranked between Control and FR. The relative area of category nr 6 of FR was higher than that obtained for the Control. HPMC addition had no impact on this category while it was improved by the addition of CMC, guar, and xanthan. The effect of xanthan on this category exceeded the results obtained for FR. Thus, only CMC and GG were similar to FR.
The relative area of category nr 11 cells was considerably higher in the FR bread than in the Control. HPMC and guar gum addition had no impact on this cell category. On the contrary, XG and CMC breads were closer to FR breads for this category of cells. These distribution results show that the Control crumb consisted mainly of gas cells of very small diameter (category nr 1). In contrast, FR crumb was more heterogeneous, with fewer cells with a diameter of 1 mm or less, and more medium (category nr 6) and particularly large cells (category nr 11). Hydrocolloid addition changed the cell distribution. HPMC addition gave breads with a homogeneous distribution of mainly very small cells (category nr 1). XG and CMC gave the closest results to French breads with greater cells size heterogeneity. Guar gum addition improved heterogeneity but to a lesser extent.
Bread dry matter results are shown in Figure 6. Dry matter content was higher in French bread (66.68%± 1.06%) than in the Control gluten-free bread (62.17%± 0.79%). However, despite this fact, sensory analyses often report a dry crumbling crumb for gluten-free breads (Gallagher and others 2003). Therefore, higher dry matter does not necessarily mean drier breads. This parameter should thus be analyzed carefully. Generally, hydrocolloid addition, except for HPMC, reduced the bread dry matter significantly. The greatest reduction in dry matter was obtained with the use of XG. This decrease could be explained by the hydrocolloids interacting with water, reducing its diffusion during baking and storage. On the contrary, HPMC addition led to breads with a significantly higher dry matter than Control breads. This result could be explained by the fact that (1) HPMC was introduced in a higher quantity than the others (2.3% for HPMC against 1.9% for xanthan gum) and thus the relative water content of the dough was slightly lower in this formulation than in the other doughs, and so the HPMC–water interactions would be masked by the initial lower water content, or (2) the quantity of HPMC added was too small to see the positive effect of this hydrocolloid on the bread water retention. HPMC can indeed be used at higher levels of supplementation in gluten-free formulations (Gujral and others 2003; Lee and Lee 2006).
The lightness of the gluten-free bread crusts varied with L* values ranging from 72.28 ± 3.27 (Control) to 79.77 ± 1.18 (HPMC), while the French bread had an L* value of 63.58 ± 2.96 (Table 3). Gluten-free breads with hydrocolloids were lighter than French breads. This difference can be attributed to the effect of hydrocolloid addition on water distribution, which impacts on Maillard browning and caramelization. The lightening of the crust color was not desirable because gluten-free breads tend to have a lighter crust color than French-style breads. XG and GG breads were not statistically different from Control breads illustrating no significant change in lightness with hydrocolloid addition. For the saturation values, French bread had a higher value than all gluten-free breads. French-style breads had a more vivid color than gluten-free breads. GG breads were statistically higher in saturation than all the other gluten-free formulations. Breads with low L* and high C* such as French-style breads had a bright tonality while gluten-free breads with high L* and low C* had a pale tonality. Based on their hue values, the breads were classified into 3 groups. The lowest hue value (73.02 ± 2.49) was obtained for French-style bread crust, that is, the overall crust shade was orange. The 2nd group was composed of gluten-free breads (except GG bread) with a hue value close to 90, suggesting an overall crust shade of yellow. GG breads were significantly different from these 2 groups with a mean hue value of 83.97 corresponding to an overall yellow–orange shade, which was the closest to the French-style bread.
Table 3—. Crust color of gluten-free breads (Control, CMC, GG, HPMC, XG, BF, EP, and WP) and French bread (FR). Superscript lowercase letters and superscript uppercase letters were used to identify significant differences in the 1st and 2nd set of trials, respectively.
As reported previously, the color and cell size distribution of breads with guar gum were close to French-style breads, therefore this hydrocolloid was chosen for the further investigations. In fact, HPMC breads had the highest specific volume and the softest crumb, but the crumb structure was made of only very small gas cells with a foamy texture and the color characteristics were not different from the Control formulation. Although GG breads had a smaller specific volume and a harder crumb than HPMC breads, they were more voluminous and softer than other formulations (basic and with other hydrocolloids).
Effects of protein sources
Guar gum was added at a concentration of 1.9% (w/w), flour basis, in every gluten-free breadmaking. For the following set of trials, GG formulation previously selected was considered as the one to improve with protein sources and as a base to quantify the effect of each protein source. The concentration of the protein source was maintained constant at 5% (w/w) but, to keep the dough hydration unchanged, adjustments were made to the amounts of flours and starches (Table 1).
The addition of egg powder and whey protein did not affect the specific volume and crumb texture of breads. On the contrary, buckwheat flour addition had an impact on both these characteristics (Figure 2 and 3). FR had a much higher specific volume than gluten-free formulations. When compared to the GG formulation, the bread specific volume was statistically increased with BF addition (3.54 ± 0.18 mL/g), while EP and WP at the same concentration had a negative impact on it. The specific volume decrease observed with whey protein addition at 5% agreed with the results of Gallagher and others (2003). Protein source addition also had a significant effect on crumb hardness, each protein source having a different impact. WP and EP breads had a harder crumb (even more for EP) than GG breads. Gallagher and others (2003) found similar increased crumb hardness for a large range of dairy proteins, which can be explained by their properties of water absorption leading to a denser crumb structure. Interestingly, BF breads (1.06 ± 0.07 N) had a softer crumb than GG breads and were closer to FR breads (0.60 ± 0.21 N).
The number of cells and the porosity of breads were unchanged after egg powder and whey protein addition (Figure 4 and 5). On the contrary, the number of cells per square centimeter decreased with buckwheat flour addition while porosity increased. The relative area occupied by gas cells was larger for BF than GG. The different formulations affected only 3 categories of cells: diameter below 1 mm (category nr 1), from 7 to 8 mm (category nr 8), and above 10 mm (category nr 11) (Table 2). All protein sources had a significant impact on the relative area of category nr 1 gas cells. For cells of category nr 8, there was no difference between GG and FR crumb, while protein sources had a slight impact, BF and WP reducing the area of this category, and EP increasing it compared to GG and FR. All breads, regardless of the protein source, had a higher relative area for category nr 11 when compared to GG, with the greatest effect obtained with buckwheat flour. Heterogeneous distribution was improved with the use of all the protein sources (reduction of the relative area of small gas cells and increase of large ones), and the best effect was observed with the use of buckwheat flour.
For bread dry matter, the BF formulation gave results close to FR breads (63.13%± 0.82% and 66.68%± 1.06%, respectively) (Figure 6). Interestingly, EP and WP breads had the same dry matter as GG breads, illustrating that egg powder and whey protein had no influence on bread dry matter.
The color characteristics are summarized in Table 3. Buckwheat flour addition improved the L* values of gluten-free breads, which decreased to 68.96 for BF compared to 73.45 for GG, therefore getting closer to French bread. Buckwheat flour was thus beneficial for obtaining a darker crust, as expected given the typical color characteristics of this flour. Protein addition had a negative effect on saturation regardless of the type of protein. Indeed, WP, EP, or BF addition decreased the C* value when compared to GG. Hue values for the crust of the gluten-free breads supplemented with a protein source varied from 80.51 to 82.89 for WP and BF, respectively, corresponding to an overall shade of yellow–orange. Nevertheless, there were statistically different h values, according to the type of protein. Buckwheat flour addition had no effect on h values, while a significant decrease was observed between GG and WP, which became closer to the shade of French bread crust. When considered individually, L*, C*, and h values had different effects on the color of gluten-free bread, so that it was difficult to establish the best formulation to obtain a bread with the same color characteristics as French bread. However, in many studies the L* value is considered to be the most important one (Gallagher and others 2003; McCarthy and others 2005) so, in this case, BF would be recommended.
The objective of this 2nd stage was to study the effects of different protein sources on gluten-free bread characteristics and also to optimize a gluten-free formulation to obtain gluten-free bread with a similar quality to French bread. The same French-style bread quality criteria as in the previous stage were used for this optimization. The addition of buckwheat flour gave the most voluminous bread, with the softest crumb and the most heterogeneous cells and with a dry matter close to that of French bread. Egg powder and whey proteins at this level of supplementation had no or limited effects on these characteristics. Thus, buckwheat flour addition gave the best bread characteristics to make bread with a similar quality to French bread.
This study identified the optimal formulation (with 1.9% guar gum [w/w] total flour basis and 5% buckwheat flour [w/w] of all flours and substitutes) for the production of gluten-free breads with French bread characteristics, suitable for coeliac patients. In addition to this physical characterization (specific volume, hardness, gas cell distribution, crust color, and bread dry matter), sensory analysis is now underway to evaluate the acceptance of this formulation by a panel of consumers. Moreover, buckwheat flour in the actual base of ingredients (rice and cornflours, corn and potato starches) was found to have interesting improving effects on the quality attributes of the bread. As the behavior of dough is directly related to the role played by the macromolecules, which determine the texture of the final product, then the proteins, starch and nonstarch carbohydrates from buckwheat merit thorough investigation.
This study was carried out with the financial support of the Commission of the European Communities, FP6, Thematic Area “Food quality and safety,” FOOD-2006-36302 EU-FRESH BAKE. The authors would like to emphasize that this article does not necessarily reflect the views of this commission and do not anticipate the commission's future policy in this area.