Analysis of the relationship between bread-making quality and dough stress during the proofing process using near-isogenic lines of ‘Harunoakebono’

Abstract

The physical properties of various white bread doughs made from the flours of ‘Harunoakebono’ and 10 genotypes of its near-isogenic lines with different compositions of high molecular weight glutenin subunit (HMWGs) were measured with the Creep method based on a Maxwell–2–element model. The expansion stress in the proofing process of various doughs was obtained by a numerical calculation method. The results indicated that doughs with high elastic characteristics, namely large relaxation time (τ₀) and regularity coefficient of viscosity (η_N), have high dough stress throughout the proofing process and high stress at the proofing end (σ_end) and conversely, the low elastic dough with the small τ₀ and η_N has the completely opposite tendency. This study also showed that there are significantly high correlations between the calculated σ_end and bread-making quality (BMQ) such as gas retention of dough and specific loaf volume (SLV). These results showed that BMQ, represented by SLV, of various white bread doughs were greatly influenced by the dough’s physical properties, especially τ₀ and η_N, which change with differences in the compositions of the HMWGs.

Introduction

It is empirically known that the bread-making quality (BMQ), such as gas retention of dough in proofing end (GRD) and specific loaf volume (SLV), are related to the physical properties of dough (PPD). Previous studies have also confirmed that the GRD and SLV of dough and bread are largely influenced by the PPD (Janssen et al. 1996, Kawai et al. 2006a, 2006b, Rao et al. 2000, Takata et al. 2000, 2003, Yamauchi et al. 2003). It has also been reported that high molecular weight glutenin subunit (HMGWs) compositions, which are wheat seed storage proteins, are closely related to PPD as well (Takata et al. 2000, 2003) and greatly affect BMQ (Krattiger et al. 1987, Lawrence et al. 1987, Takata et al. 2000, 2003). They are encoded by the Glu-1 loci on the long arms of group 1 chromosomes (Payne et al. 1980). The alleles at these three loci have different effects on BMQ. However, low molecular weight glutenin subunit (LMWGs) and gliadin, which are also wheat seed storage proteins except for HMWGs, are also related to the PPD (Benedettelli et al. 1992, Gupta et al. 1994, Ito et al. 2011, Jin et al. 2013, Zhang et al. 2012).

In addition, lipids and pentosans seem to be related to BMQ (Morrison et al. 1989, Roels et al. 1993). Previous studies have found that HMWGs compositions are a major factor affecting BMQ of wheat flour dough, but there are many other factors that affect BMQ, such as wheat flour characteristics. Therefore, to exclude factors other than the compositions of HMWGs that might have an impact on bread-making properties, ‘Harunoakebono’ (HA) and 10 genotypes of its near-isogenic lines (NILs), whose genetic background except for the differences in HMWGs compositions is almost the same, were used in these experiments. By using the NILs, it is possible to clearly analyze the relationship between the differences in HMWGs compositions (differences in PPD) and their impact on BMQ.

Previous studies have been done on the relationship between PPD and BMQ, but the behavior of the PPD during the proofing process of bread-making has not been extensively studied. Therefore, it is not clear how the PPD just after preparation affect its behavior in the proofing process and how GRD and SLV change by its behavior.

Bloksma (1957) studied the mechanism of dough expansion with an Alveographe, which measures the expansion of dough, using a Maxwell-2-element model (Maxwell model) and calculated stress behavior during dough expansion using numerical analysis. He reported that the stress relaxation time was the most important factor for the expansion ability. Matsumoto (1981) also studied the stress of dough in the fermentation process theoretically using the Maxwell model under the condition of constant cross-section and nominal strain. The study showed that stress during dough expansion changes qualitatively depending on the initial stress. Although these papers explained the stress behavior in the expansion process to some extent, they were not applied to the actual bread making process. These previous studies did not also examine the influence of the initial PPD on the behavior of PPD in the actual expansion process of bread-making. In addition, the correlation between the final stress (σ_end) of PPD in expansion process and BMQ has not been analyzed.

In this study, dough samples without yeast with various PPDs were prepared using flours of HA and 10 genotypes of its NILs with different HMWGs compositions. The PPD was expressed with a Maxwell model and the coefficients of physical properties of these various doughs were measured by the Creep method. Subsequently, the changes in various doughs stresses in the proofing process were simulated by numerical analysis, assuming that the PPD did not change in the bread-making process, using the initial PPD data. Based on the obtained results, the relationship between the simulation results and the actual GRD and SLV was analyzed in detail. Specifically, we analyzed the correlation between σ_end and BMQ such as GRD and SLV and examined how the initial PPD affects σ_end. Finally, the effect of HMWGs compositions on BMQ was discussed.

Materials and Methods

Analytical model of PPD

The deformation model of dough for the numerical analysis is shown in Fig. 1. It was assumed that the dough is an aggregate of many cylindrically shaped doughs (left side in Fig. 1) and that the dough is stretched in a longitudinal direction by gas inflation at a constant expanding speed of C (m/s) with time t (s), but the total volume is kept constant during the expansion. The initial height of dough are defined as a₀ (m) (right side in Fig. 1). From the measurements of gassing power (GP) of actual doughs, the GP of carbon dioxide that the yeast generates is almost constant. Then, assuming that little gas is lost from the doughs, the expansion rate of each dough was set as constant. As will be described later, the expansion rate (C (m/s)) of the dough was obtained from the GP of dough and the average area of the bread pan.

Fig. 1

Dough expansion model for dough stress simulation during the proofing process. a₀: initial height of dough, C: expansion rate of dough, t: time.

By using the data of the actual dough strain rate of longitudinal direction, the strain rate of dough (γ_(t) (1/s)) at t (s) is defined as Eq. (1).

  

$${\dot{\mathrm{\gamma }}}_{(\text{t})}=\frac{1}{{\text{a}}_{(\text{t})}}\frac{{\text{da}}_{\text{(t})}}{\text{dt}}=\frac{\text{C}}{{\text{a}}_0+\text{Ct}}$$Eq. (1)

Here, a_(t) (m) is the height of the dough at t (s). From this equation, γ_(t) is not constant although the expanding speed of the dough is constant.

Next, the Maxwell model was used to represent the PPD as indicated in Fig. 2. In this figure, E₀ (Pa), η_N (Pa·s), γ₀ (−), and γ_N (−) are the instantaneous elasticity, the regularity coefficient of viscosity, the strain of instantaneous elasticity region and the strain of regularity coefficient viscosity region, respectively. In addition, the relaxation time (τ₀ (s)) is defined as τ₀ = η_N/E₀. The strain rate of dough (γ_(t)) is represented using a Maxwell model as Eq. (2).

Fig. 2

Maxwell model for analysis of physical properties of dough. E₀: instantaneous elasticity, η_N: regularity coefficient of viscosity, γ₀: strain of instantaneous elasticity region, γ_N: strain of regularity viscosity region.

  

$${\dot{\mathrm{\gamma }}}_{(\text{t})}=\frac{1}{{\text{E}}_0}\frac{\text{d}{\mathrm{\sigma }}_{(\text{t})}}{\text{dt}}+\frac{{\mathrm{\sigma }}_{(\text{t})}}{{\mathrm{\eta }}_{\text{N}}}$$Eq. (2)

Here, σ_(t) (Pa) is the stress of dough at t during expansion. Equation (3) is obtained by substituting the right side of Eq. (1) for the left side of Eq. (2) and arranging the equation. Equation (3) was obtained to show the stress changes in the expansion process of dough.

  

$$\frac{\text{d}{\mathrm{\sigma }}_{(\text{t})}}{\text{dt}}={\text{E}}_0\frac{\text{C}}{{\text{a}}_0+\text{Ct}}-\frac{{\text{E}}_0}{{\mathrm{\eta }}_{\text{N}}}{\mathrm{\sigma }}_{(\text{t})}$$Eq. (3)

Here, σ₍₀₎ (Pa) is the initial stress, which was set as 0 Pa in the numerical analysis of σ_(t). The σ_(t) of several doughs with various physical properties was simulated with Eq. (3) by Euler’s method.

The main reasons for using the Maxwell model as the PPD model in this research are as follows. (1) Previous reports show that the behavior of PPD to changes during the expansion process can be represented by this model (Matsumoto 1981). (2) This model is simple and it is easy to numerically calculate the stress changes of dough in the expansion process (Bloksma 1957, Kawai et al. 2006a). (3) In the expansion process of dough at a very slow strain rate, it is reasonable that the retardation elasticity region, which is normally present in the condition of rapid deformation of dough, is included in the instantaneous elasticity region.

Plant and flour materials

The NILs substituted for single and double HMWGs were used in this study, which were developed by Takata et al. (2000, 2003). The NILs were developed by crossing the recurrent parent, HA, and the donor parents, Haruyutaka, Norin 61, Takunekomugi and Chihokukomugi (Table 1). The names of various HMWGs encoded by Glu-1 loci in Table 1 were the same as those of the report of Takata et al. (2003), which is based on the band position of SDS-PAGE. Takata et al. (2000, 2003) reported that HMWGs 1 has almost the same properties as HMWGs 2* of HA and has an intermediate effect on the improvement of the strength of dough, that HMWGs 17 + 18, 7 + 8, 6 + 8 and 20 have a slightly stronger, equal, slightly weaker, and very weakening effect on dough strength compared to HMWGs 7 + 9 of HA, respectively and that compared to HMWGs 5 + 10 of HA, HMWGs 2 + 12, 2.2 + 12 and 4 + 12 have the effect of significantly reducing the strength of the dough.

Table 1 HMWGs compositions of recurrent parent and donor parents

Cultivar	HMWGs
	Glu-A1	Glu-B1	Glu-D1
HA	2*	7 + 9	5 + 10	recurrent parent
Haruyutaka	1	17 + 18	2 + 12	donor parent
Norin 61	2*	7 + 8	2.2 + 12	donor parent
Takunekomugi	1	6 + 8	4 + 12	donor parent
Chihokukomugi	1	20	2 + 12	donor parent

HMWGs: high molecular weight glutenin subunit, HA: Harunoakebono. The underline shows HMWGs introduced to HA.

The homozygous genotypes of the NILs were checked for glutenin by SDS-PAGE (Nakamura et al. 1990) and for gliadin components by acid polyacrylamide gel electrophoresis (A-PAGE) (Bushuk and Zillman 1978). Eight kinds of NILs substituted for single HMWGs, two kinds of NILs for double HMWGs, and HA, a recurrent parent, were cultivated according to the conditions reported by Yamauchi et al. (2007).

The obtained grain samples to which water was added so as to make their moisture content 16% were milled with a Bühler test mill (Bühler Inc., Uzwil, Switzerland) and flours of 60% extraction rate (flours) were obtained. The protein and ash contents (w/w, 13.5% moisture base) were respectively measured using a near-infrared reflectance instrument (Inframatic 8120, PerCon Co., Hamburg, Germany) and the method of the American Association of Cereal Chemists (AACC) as reported by Yamauchi et al. (2014).

Dough preparations and bread-making tests

The bread-making tests were done using the no-time method and standard white bread formulation. The tests were partially modified from the method reported by Yamauchi et al. (2001). Namely, the proofing temperature in this study was changed from 38 to 30°C. The proofing process was also done until the top height of the dough reached 1 cm from the top edge of the pan.

The optimum amounts of water for dough preparations and bread-making tests were determined with a Farinograph at 500 BU according to the AACC method (1991). The optimum water absorption of HA and NILs flours were 59.4 to 62.0% for dough preparations and bread-making tests.

The GRD was evaluated by measuring the maximum expansion volume of 20 g of dough proofed at 30°C and 85% relative humidity in a cylinder under low pressure, following steps outlined by Yamauchi et al. (2000).

The GP of 20 g of dough after bench time was measured at 30°C for 3 hrs at 5 min intervals using a Fermograph II (ATTO Co., Ltd., Tokyo, Japan) as reported by Santiago et al. (2015). The GP rates of each 100 g of dough used for bread making in the proofing process were calculated by assuming that the GP rates are constant. Using the GP data, the vertical expansion rates in the proofing process of various doughs were calculated to take into consideration the volume of 100 g of dough after molding and the average area of a baking pan, assuming that the shape of dough after molding was a rectangular parallelepiped.

The SLV of various breads cooled at room temperature for 1 h after baking was measured by the AACC International (AACCI) rapeseed-displacement method (2000). Individual slices from the breads were photocopied using a copy machine to evaluate the size and crumb grain of each bread.

Measuring conditions of PPD

The measurements of PPD were established using the doughs just after mixing. The doughs were divided into 30 g pieces, rounded to a spherical shape, compressed to 1.5 cm thickness using a flat box, put inside a polyethylene bag to prevent water evaporation, and then kept in an incubator at 30°C for 60 min to ensure a flat surface and the stress relaxation of the dough. The actual PPD of various samples were measured via the Creep method using a Rheoner (Model RE33005, Yamaden Co., Ltd., Tokyo, Japan) as reported by Kawai et al. (2006a). The load for the Creep measurement was 499 Pa. This value was determined to accurately measure the physical property values of various doughs within the range where the relationship between the load and the strain is almost linear. Each coefficient, E₀ and η_N, of the Maxwell model was determined with the automatic analytical software of the Rheoner. Each τ₀ was calculated using the PPD coefficients.

Simulation of dough stress during the proofing process

The simulations of the stresses of various doughs during the proofing process were carried out using Eq. (3) by Euler’s method. The boundary and initial conditions for the simulations were σ₍₀₎ = 0 Pa, a₀ = 1.4 × 10⁻² m, and C at 0 s = 6.50 × 10⁻⁶ m/s, respectively. Since there was almost no difference in GP in all doughs, the average value thereof was used as the C value in all stress simulations. The calculating interval used in Euler’s method was 20 s. The value of a₀ was calculated using the volume (98.1 ml) of 100 g of dough just after molding and the average area (68.3 cm²) of the baking pan, assuming that the dough was a rectangular parallelepiped.

Statistical analysis

All experimental data are shown as mean ± SD, except for τ₀. Significant differences of the data in Tables 2 and 3 were evaluated using the analysis of variance at 5% level of p-value and Tukey’s multiple range test using Excel statistical software 2012 (SSRI Corporation, Tokyo, Japan). Correlation analyses between various data were performed using Microsoft Excel 2012 software.

Table 2 HMWGs compositions, flour protein content, and ash content of HA and NILs of Glu-1 alleles

Flour samples	HMWGs			Properties of flours
	Glu-A1	Glu-B1	Glu-D1	Protein (%)	Ash (%)
HA	2*	7 + 8	5 + 10	10.1 ± 0.2cd	0.39 ± 0.00a
NIL1	1	7 + 9	5 + 10	10.4 ± 0.2abc	0.38 ± 0.01a
NIL6 + 8	2*	6 + 8	5 + 10	9.9 ± 0.2d	0.41 ± 0.01a
NIL7 + 8	2*	7 + 8	5 + 10	9.3 ± 0.1e	0.39 ± 0.02a
NIL17 + 18	2*	17 + 18	5 + 10	10.1 ± 0.1cd	0.41 ± 0.02a
NIL20	2*	20	5 + 10	10.2 ± 0.0bcd	0.39 ± 0.01a
NIL4 + 12	2*	7 + 9	4 + 12	10.5 ± 0.4abc	0.39 ± 0.01a
NIL2 + 12	2*	7 + 9	2 + 12	10.7 ± 0.2a	0.38 ± 0.01a
NIL2.2 + 12	2*	7 + 9	2.2 + 12	10.4 ± 0.1abc	0.39 ± 0.01a
NIL17 + 18/2.2 + 12	2*	17 + 18	2.2 + 12	10.6 ± 0.1ab	0.41 ± 0.01a
NIL20/2.2 + 12	2*	20	2.2 + 12	10.3 ± 0.1abcd	0.40 ± 0.02a

HMWGs: high molecular weight glutenin subunit, HA: Harunoakebono, NILs: near-isogenic lines, NIL: near-isogenic line. The letters of underline shows substituted HMWGs. Data are shown in mean ± SD (n = 3). Protein and ash contents are based on 13.5% moisture content. The values followed by different letters within column are significantly different (p < 0.05). The analysis of variance between the data was evaluated using Tukey’s multiple range test of Excel statistical software 2012.

Table 3 Physical properties of various bread doughs

Flour samples	E0 (kPa)	ηN(Pa·s × 105)	τ0 (s × 102)
HA	3.64 ± 0.32ab	10.96 ± 2.31abc	3.01
NIL1	3.16 ± 0.35bcd	11.85 ± 2.42ab	3.76
NIL6 + 8	3.60 ± 0.76abc	12.48 ± 2.30a	3.46
NIL7 + 8	3.95 ± 0.28a	13.24 ± 1.90a	3.35
NIL17 + 18	3.29 ± 0.33bcd	12.88 ± 3.02a	3.94
NIL20	3.08 ± 0.34bcd	8.39 ± 1.40cd	2.72
NIL4 + 12	3.60 ± 0.51abc	8.80 ± 1.48c	2.44
NIL2 + 12	3.28 ± 0.39bcd	8.37 ± 2.29cd	2.55
NIL2.2 + 12	2.98 ± 0.23de	5.88 ± 0.37de	1.98
NIL17 + 18/2.2 + 12	3.05 ± 0.47cd	9.39 ± 0.99bc	3.07
NIL20/2.2 + 12	2.04 ± 0.30e	4.66 ± 0.17e	1.94

E₀: instantaneous elasticity, η_N: regularity coefficient of viscosity, τ₀: relaxation time, HA: Harunoakebono, NIL: near-isogenic line. Data are shown in mean ± SD (n = 10–12). The physical properties of dough were measured using no yeast doughs. The values followed by different letters within column are significantly different (p < 0.05). The analysis of variance between the data was evaluated using Tukey’s multiple range test of Excel statistical software 2012.

Results

Characteristics of HA and NILs flours

HMWGs compositions, flour protein content and ash content are shown in Table 2. Except for the introduced subunits, these HA and NILs showed the same electrophoresis pattern by SDS-PAGE and the same gliadin bands by A-PAGE (data not shown). All samples, including HA, had hard grain. The flour protein and ash contents ranged 9.3 to 10.7% and 0.38 to 0.41%, respectively. The protein contents of partial NILs (NIL7 + 8, NIL2 + 12 and NIL17 + 18/2.2 + 12) were significantly different from that of HA. Especially, the protein content of NIL 2 + 12 was significantly higher than that of HA and NIL 7 + 8 was significantly lower than HA. However, the range of protein contents of HA and its NILs was 1.4% and not much different. On the other hand, the ash contents were not significantly different among all the samples and the range was also small. The GP from the bread dough did not significantly differ among all samples (data not shown).

Physical properties of various doughs

The doughs for the measurements of various PPDs were prepared using the HA and NILs flours shown in Table 2, which were produced by the same way as those used in the bread-making tests except that no yeast was added. Because when leavened doughs are used, PPD measurements are less accurate. The PPD of HA and its NILs are shown in Table 3. The E₀, η_N, and τ₀ values ranged from 2.04 to 3.95 kPa, 4.66 × 10⁵ to 13.24 × 10⁵ Pa·s and 1.94 × 10² to 3.94 × 10² s, respectively. The E₀ and η_N values of partial NILs (E₀: NIL2.2 + 12, NIL17 + 18/2.2 + 12 and NIL20/2.2 + 12; η_N: NIL2.2 + 12 and 20/2.2 + 12) were significantly different from the HA. Due to the difference of HMWGs compositions, the τ₀ value drastically changed as the E₀ and η_N values, especially the latter, of PPD changed. In terms of the η_N and τ₀, the NILs with HMWGs 5 + 10 except for NIL 20 (NILs with 5 + 10) showed clearly larger η_N and τ₀ values than the other NILs without HMWGs 5 + 10 except for NIL 17 + 18/2.2 + 12 (NILs without 5 + 10). In addition, NIL 20 and NIL 17 + 18/2.2 + 12 showed a value approximately midway between NILs with 5 + 10 and NILs without 5 + 10, and a value close to the NILs without 5 + 10 or the NILs with 5 + 10, respectively. NIL 20/2.2 + 12 also showed the lowest PPD, very low η_N and τ₀ values, among all samples. NIL 17 + 18 showed large values of η_N and τ₀ due to double strengthening effects of HMWGs 17 + 18 and 5 + 10, and NIL 17 + 18 also showed the largest τ₀ value among all samples.

Simulation results of dough stresses

The simulations of the stresses of various doughs during the proofing process were done by using the values of each coefficient of the various PPDs shown in Table 3. The simulation results are shown in Figs. 3 and 4. The proofing process on bread-making was done until the top height of the dough reached 1 cm from the top edge of the pan and the time was measured. And the each value of σ_end is the respective simulation dough stress in the each proofing end time. These results indicate that all stress curves sharply increased at the initial stage of expansion and then eventually decreased.

Fig. 3

The stress simulation of doughs of various physical properties in the proofing process. HA: Harunoakebono, NIL: near-isogenic line, σ_end: dough stress in end of proofing process.

Fig. 4

The stress simulation of doughs of various physical properties in the proofing process. HA: Harunoakebono, NIL: near-isogenic line, σ_end: dough stress in end of proofing process.

Fig. 3 shows the simulation results for the dough stresses in the expansion process of NILs substituted for HMWGs of Glu-A1 and Glu-B1 alleles. The simulation results of dough stresses of all NILs except for NIL 20 did not differ much from HA. On the other hand, NIL 20 showed a considerably smaller stress peak and σ_end value than the HA.

Fig. 4 shows the simulation results of the dough stresses in the expansion process of NILs substituted for HMWGs of Glu-B1 and Glu-D1 alleles. Regarding the HA and its NILs in Fig. 4, the stress peaks and σ_end of all NILs showed lower values than the HA. In particular, those of NIL 2.2 + 12 and NIL 20/2.2 + 12, especially the latter, showed the very low values compared to the other NILs.

Relationship between various PPDs, GRD and SLV

It was found that PPD changed substantially with substitutions of HMWGs (Table 3 and Figs. 3, 4) and the dough stress in the proofing process also decreased largely due to the weakening of the PPD (mainly η_N and τ₀). The bread-making tests were carried out using the HA and NILs flours shown in Table 2 and the BMQ data such as GRD and SLV were obtained. The correlations between PPD, GRD and SLV were analyzed. The results are shown in Table 4. All PPD were correlated significantly with GRD and SLV in a logarithmic approximation and the correlations between PPD except for E₀, GRD and SLV were extremely high and significant at a less than 0.1% level of p-value.

Table 4 Correlation coefficients between PPD, GRD and SLV

	GRD (ml)	SLV (ml/g)
E0 (kPa)	0.853***	0.646*
ηN (Pa·s × 105)	0.951***	0.904***
τ0 (s × 102)	0.851***	0.902***
σend (daPa)	0.948***	0.916***

The correlation coefficients are a logarithmic approximation coefficient.

PPD: physical properties of dough, GRD: gas retention of dough, SLV: specific loaf volume, E₀: instantaneous elasticity, η_N: regularity coefficient of viscosity, τ₀: relaxation time, σ_end: dough stress in end of proofing process, *,***: significant level; p < 0.05 and p < 0.001, respectively.

Fig. 5 shows the correlations between σ_end, GRD and SLV, which have comprehensively the highest correlation coefficients in Table 4. It can be seen that GRD and SLV can mostly be estimated by the value of σ_end.

Fig. 5

Correlations between σ_end, GRD and SLV. The lines are log correlation curves. GRD: gas retention of dough, SLV: specific loaf volume, σ_end: dough stress in end of proofing process, r: correlation coefficient, ***: significant level (p < 0.001), y: GRD or SLV, x: σ_end.

In order to clarify PPD which greatly affects σ_end value, the correlations between σ_end and various PPDs were analyzed. Table 5 shows the correlation coefficients between σ_end and various PPDs values. It was found that the correlations between σ_end, η_N and τ₀ are extremely high and that σ_end is mostly determined by the η_N and τ₀ values of the doughs in this study.

Table 5 Correlations coefficients between σ_end and various values of physical properties

	E0 (kPa)	ηN (Pa·s × 105)	τ0 (s × 102)
σend (daPa)	0.736**	0.997***	0.946***

The correlation coefficients are a coefficient of linear approximation.

E₀: instantaneous elasticity, η_N: regularity coefficient of viscosity, τ₀: relaxation time, σ_end: dough stress in end of proofing process, **,***: significant level; p < 0.01 and p < 0.001, respectively.

Fig. 6 shows the correlation between GRD and SLV. There was a significant correlation between GRD and SLV at less than 0.1% level of p-value. This indicates that GRD has a great influence on SLV.

Fig. 6

Correlation between GRD and SLV. The line is a straight line of linear correlation. GRD: gas retention of dough, SLV: specific loaf volume, r: correlation coefficient, ***: significant level (p < 0.001), y: SLV, x: GRD.

Evaluation of size, crumb grain and SLV of breads

The photocopies and SLV of the breads obtained from 11 HA and its NILs are shown in Fig. 7. The biggest bread was obtained from the dough of NIL 17 + 18 and the bread obtained from the dough of NIL20/2.2 + 12 was the smallest. Fig. 7 also shows that the NILs with 5 + 10 including NIL20, especially NIL17 + 18, had the largest SLV and the best crumb grain and that the NILs without 5 + 10, especially NIL 20/2.2 + 12, had a low SLV and rougher and non-uniform crumb grain. On the other hand, NIL 17 + 18/2.2 + 12 showed a large SLV and relatively good crumb grain among the NILs without 5 + 10, which had characteristics similar to those of HA.

Fig. 7

Size and crumb grain images and SLV of breads made from doughs of HA and NILs. HA: Harunoakebono, NIL: near-isogenic line, NILs: near-isogenic lines, SLV: specific loaf volume.

Fig. 7 also showed that the crumb grain of NILs having weak elastic properties (low τ₀), especially that of NIL20/2.2 + 12, exhibited crumb grain with very non-uniform and large bubbles, which indicates that the bubbles in the dough broke and coalesced in the proofing process.

Discussion

Characteristics of flours and physical properties of doughs

Table 2 showed that the protein content of HA and its NILs were somewhat different and that the ash content and GP were almost the same among all samples. These results suggest that the differences of protein and ash contents of flours and GP of doughs among the HA and the NILs hardly influence the differences in PPD and BMQ and that the flour compositions of all samples except for HMWGs compositions were nearly same. Table 3 showed that the PPD of HA and the NILs, especially η_N and τ₀, greatly change due to the difference of HMWGs compositions. The values of η_N and τ₀ are indicators of the elasticity of PPD and the larger values show that the dough has the stronger elastic properties. The values of η_N and τ₀ of HA and its NILs based on the Maxwell model in this study showed a similar tendency compared to the breaking forces of doughs measured with the same HA and its NILs by Takata et al. (2000, 2003). They also examined the influence of various HMWGs on the PPD and reported that HMWGs 5 + 10 and 17 + 18 had the largest and second largest effects on strengthening of the PPD, respectively and that HMWGs 20, 4 + 12, 2 + 12 and 2.2 + 12 had very strong effect to weaken the elastic properties of dough. The data in Table 3 were mostly consistent with the results reported by Takata et al. (2000, 2003).

In addition, much knowledge has been accumulated about the influence of differences in HMWGs compositions on the physical properties of dough (gluten), which is because research on the alleles of three loci (Glu-A1, Glu-B1 and Glu-D1 locus) of HMWGs has progressed rapidly. Basically, HMWGs, except for HMWGs encoded by Glu-A1 locus, are composed of two subunits of x-type and y-type encoded by two genes in each locus (Payne 1987). It is known that the y-type subunit has more cysteine residues related to the disulfide bond in its molecular than the x-type subunit (Megan and Skerritt 1999, Morel and Bonicel 1996). Models have also been proposed for glutenin polymers composed of HMWGs and LMWGs that are believed to form a gluten backbone. In these models, HMWGs, especially y-type HMWGs, play a major role in the polymerization of glutenin, and LMWGs bind as branches to the backbone of HMWGs polymers (Shewry et al. 2002, Wieser 2007). In these structural models, y-type HMWGs can bind to many HMWGs molecules using many cysteine residues. On the other hand, it is thought that the contribution of x-type HMWGs to the glutenin polymerization in these models is limited, because most x-type HMWGs have only two cysteine residues related to the interchange disulfide bond (Pirozi et al. 2008, Shewry et al. 2003, Wieser 2007). However, only HMWGs 5 among the x-type HMWGs have three cysteine residues related to interchange disulfide bond and it is known that the contribution to glutenin polymerization is specifically large. On the other hand, x-type HMWGs 20 have only two cysteine residues in the molecule (The other x-type HMWGs except for HMWGs 5 usually have four cysteine residues), so it is believed that the contribution to glutenin polymerization is very small (Megan and Skerritt 1999, Morel and Bonicel 1996, Pirozi et al. 2008). Furthermore, it is also reported that the polymerization of glutenin polymers significantly enhances the physical properties of the dough (Orth and Bushuk 1972). From these findings, it is seemed that the remarkable strengthening effect of the dough’s physical properties by HMWGs 5 + 10 and the strong weakening effects by HMWGs 20, 4 + 12, 2 + 12 and 2.2 + 12 in this study can be reasonably explained.

Simulation behaviors of dough stresses

Figs. 3 and 4 indicated that all stress curves of doughs of HA and its NILs sharply increased at the initial stage and then decreased. The reason is that the height of the dough is low during the initial stage, the strain rate is high and the increase of stress occurs more rapidly than the relaxation of stress in the initial stage. On the other hand, it seems that the decrease in stress after the peak is because as the height of the dough increases with the fermentation period and causes a decrease in the strain rate, the relaxation of dough stress becomes dominant compared to the increase of stress.

Fig. 3 showed the simulation results for the dough stresses of NILs substituted for HMWGs of Glu-A1 and Glu-B1 alleles. As described above, HMWGs 5 + 10 coded by Glu-D1 allele shows the largest reinforcing effect of the PPD among all HMWGs. Since all HA and its NILs shown in Fig. 3 have this HMWGs 5 + 10, the simulation results of dough stresses of all NILs except for NIL 20 did not differ much from HA. On the other hand, NIL 20 showed a considerably smaller stress peak and σ_end value than the HA. This is considered to be related to that HMWGs 20 coded by Glu-B1 allele as well as HMWGs 4 + 12, 2 + 12 and 2.2 + 12 coded by Glu-D1 allele shows the largest weakening effect on the PPD among all HMWGs. In terms of the substitutions of HMWGs of Glu-A1 and Glu-B1 alleles, these results suggest that the substitution of only HMWGs 20 greatly affected the PPD of the NIL and the substitution of other HMWGs had little effect on the PPD of NILs.

Fig. 4 showed the simulation results of the dough stresses of NILs substituted for HMWGs of Glu-B1 and Glu-D1 alleles. The stress peaks and σ_end of all NILs showed lower values than the HA, because all NILs did not have HMWGs 5 + 10. The particular low values of NIL 2.2 + 12 and NIL 20/2.2 + 12, especially the latter, seem to be related to HMWGs 20 coded by Glu-B1 allele and HMWGs 2.2 + 12 coded by Glu-D1 allele. HMWGs 2.2 + 12 had the largest weakening effect on the PPD among all of the HMWGs of NILs in Fig. 4. NIL 20/2.2 + 12 also showed extremely weak PPD due to double weakening effect of HMWGs 20 and 2.2 + 12 and, as a result, had very low peak stress and σ_end.

Relationship between various PPDs, GRD and SLV

Table 4 showed that the correlations between PPD except for E₀, GRD and SLV are extremely high. Since η_N, τ₀ and σ_end are the PPD that relate highly to the elastic properties of dough, those results show that good bread having large GRD and SLV, which are indicators of BMQ, can be obtained from dough exhibiting more elastic physical properties.

Fig. 5 showed the high correlations between σ_end, GRD and SLV. These results indicated that the more elastic dough, in which the overall dough stress and σ_end in the proofing process are higher, shows the higher GRD and that the bread of larger SLV is obtained from the dough. These results basically corresponded to reports by Bloksma (1957) and Matsumoto (1981), which showed that dough exhibiting more elastic properties had better expansion characteristics and BMQ (mainly the size of SLV).

Table 5 showed that the correlations between σ_end, η_N and τ₀ are extremely high. Both values of η_N and τ₀ are physical properties that relate to the elastic properties of dough. When those values are high, it is more likely to have elevated dough stress than the relaxation of the stress during the dough expansion process. Therefore, it is reasonable that the correlations between σ_end, η_N and τ₀ are high. In addition, although the PPD value that most influences the rise and relaxation of the stress in the dough expansion process is generally considered to be τ₀, the correlation coefficient between σ_end and η_N was higher than that between σ_end and τ₀ (Table 5). This seems to be related to the fact that the E₀ values of PPD in HA and its NILs do not show a large change, but the change of the value of η_N is large. It seems that the main cause is that the η_N values consequently varied in conjunction with τ₀. These results are basically consistent with those reported by Bloksma (1957) that stress relaxation time greatly affects dough expansion stress.

Fig. 6 showed that the very high correlation between GRD and SLV. Yamauchi et al. (2000) reported that there was a high correlation between GRD and SLV. However, the correlation between GRD and SLV has not been studied using HA and its NILs flours, which are almost identical in genetic background except for HMWGs compositions. Therefore, they did not find the high correlation between GRD and SLV that can be seen in Fig. 6. This suggests that the high correlation between GRD and SLV in Fig. 6 is mainly caused by that the factors except for the differences of PPD do not almost affect this correlation, which was only influenced by the differences in HMWGs compositions. It was also found that when using the doughs like HA and its NILs, the SLV of the breads is mostly determined by the value of GRD, which is related to the differences in the PPD.

In order to increase the value of this GRD, it is so important to make the PPD more elastic by optimizing the HMWGs compositions (Figs. 3–6 and Tables 4, 5). Namely, the values of η_N and τ₀ of the PPD are improved by optimizing the HMWGs compositions and, thereby, the σ_end value increases. Since σ_end is highly correlated with GRD and SLV, better bread having higher SLV is obtained from dough with a higher σ_end value, meaning that the dough has more elastic properties. The HMWGs showing the greatest positive effect on this increase in the σ_end value (improvement of the PPD) was 5 + 10 and followed by 17 + 18.

Overall BMQ of various doughs of HA and NILs

Fig. 7 showed that the size and crumb grain images and SLV of the breads made from 11 HA and its NILs doughs. These results mostly corresponded to those of the bread-making tests using the same HA and NILs reported by Takata et al. (2000, 2003). The photocopies of NILs having weak elastic properties (low τ₀), especially that of NIL20/2.2 + 12, exhibited the crumb grain with very non-uniform and large bubbles. These results were mostly consistent with those reported by Matsumoto (1981, 1991). He reported that when a soft dough with weak elastic properties (low τ₀) expands with yeast fermentation, the bubble membrane in the dough breaks and coalesces during the expansion process.

From the data of the experiments using HA and its NILs doughs with various PPD, it was observed that the σ_end value obtained by simulation significantly correlated with the GRD and SLV. Therefore, to obtain a bread having a large SLV and good crumb grain, it is necessary to keep dough stress high and suppress the coalescence of gas bubbles in the dough during the proofing process. BMQ, such as GRD and SLV, are also strongly related to the PPD, especially η_N and τ₀. Among HA and its NILs, NIL 17 + 18 having the most overall elastic PPD showed the best BMQ and NIL 20/2.2 + 12 which had the completely opposite PPD showed the lowest BMQ.

In this study, the HA and its NILs flours with a mostly homogeneous background except for HMWGs compositions were used, and the preparation of the various doughs was strictly controlled and bread-making conditions except for the HMWGs compositions of the flours were all performed under the same conditions. Therefore, the data concerning BMQ in this study are more reliable than those of normal bread-making tests. It is expected to study whether similar results can be obtained using different bread-making conditions and dough compositions with the HA and its NILs flours used in this experiment.

Author Contribution Statement

K. T. and H. Y. conceived and planned this study, and performed the analytic calculations and the numerical simulations. K. M., D. G. and T. N. carried out the experiments. K. T. and H. Y. also took the lead in writing the manuscript.

Literature Cited

• AACC (1991) Approved Methods of the AACC, 8th ed. Method 08-01. The American Association of Cereal Chemists, St. Paul, MN.
• AACCI (2000) Approved Methods of the AACCI, 11th ed. Method 10-05.01. The American Association of Cereal Chemists International, St. Paul, MN.
•  Benedettelli,  S.,  B.  Margiotta,  E.  Porceddu,  M.  Ciaffi and  D.  Lafiandra (1992) Effects of the lack of protein controlled by genes at the Gli D1/Glu-D3 loci on the breadmaking quality of wheat. J. Cereal Sci. 16: 69–79.
•  Bloksma,  A.H. (1957) A calculation of the shape of the Alveograms of some rheological model substances. Cereal Chem. 34: 126–136.
•  Bushuk,  W. and  R.R.  Zillman (1978) Wheat cultivar identification by gliadin electrophoregrams. I. Apparatus, method and nomenclature. Can. J. Plant Sci. 58: 505–515.
•  Gupta,  R.B.,  J.G.  Paul,  G.B.  Cornish,  G.A.  Palmer,  F.  Bekes and  A.J.  Rathjen (1994) Allelic variation at glutenin subunit and gliadin loci, Glu-1, Glu-3 and Gli-1, of common wheats. I. Its Additive and interaction effects on dough properties. J. Cereal Sci. 19: 9–17.
•  Ito,  M.,  S.  Fushie,  W.  Maruyama-Funatsuki,  T.M.  Ikeda,  Z.  Nishio,  K.  Nagasawa,  T.  Tabiki and  H.  Yamauchi (2011) Effect of allelic variation in three glutenin loci on dough properties and bread-making qualities of winter wheat. Breed. Sci. 61: 281–287.
•  Janssen,  A.M.,  T.  van Vliet and  J.M.  Vereijken (1996) Fundamental and empirical rheological behaviour of wheat flour doughs and comparison with bread making performance. J. Cereal Sci. 23: 43–54.
•  Jin,  H.,  Y.  Zhang,  G.  Li,  P.  Mu,  Z.  Fan,  X.  Xia and  Z.  He (2013) Effects of allelic variation of HMW-GS and LMW-GS on mixograph properties and Chinese noodle and steamed bread qualities in a set of Aroona near-isogenic wheat lines. J. Cereal Sci. 57: 146–152.
•  Kawai,  H.,  F.  Tanaka,  H.  Takahashi and  H.  Yamauchi (2006a) Measurement of physical properties and expansion ability of dough for bread making. Food Preservation Science 32: 209–216.
•  Kawai,  H.,  F.  Tanaka,  H.  Takahashi,  N.  Hashimoto and  H.  Yamauchi (2006b) Relationship between physical properties of dough and expansion ability during bread-making. Food Sci. Technol. Res. 12: 91–95.
•  Krattigar,  A.F.,  P.I.  Payne and  C.N.  Law (1987) The relative contribution of proteins and other components to breadmaking quality of varieties determined using chromosome substitution lines. In:  Laszity,  R. and  F.  Bekes (eds.) Proc. 3rd Int. workshop gluten proteins. Budapest, Hungary. World Scientific, Singapore. pp. 254–265.
•  Lawrence,  G.J.,  H.J.  Moss,  K.W.  Shepherd and  C.W.  Wrigley (1987) Dough quality of biotypes of eleven Australian wheat cultivars that differ in high-molecular-weight glutenin subunit composition. J. Cereal Sci. 6: 99–101.
•  Matsumoto,  H. (1981) Various basic research on the expansion of bread. Cyourikagaku 68: 215–221.
•  Matsumoto,  H. (1991) Basic research on expansion of dough and loaf Internal pressure in yeast dough II. Bulletin of Kobe Women’s University 24: 161–178.
•  Megan,  P.L. and  J.H.  Skerritt (1999) The glutenin macropolymer of wheat flour doughs: Structure-function perspectives. Trends Food Sci. Technol. 10: 247–253.
•  Morel,  M.-H. and  J.  Bonicel (1996) Determination of the number of cysteine residues in high molecular weight subunits of wheat glutenin. Electrophoresis 17: 493–496.
•  Morrison,  W.R.,  C.N.  Law,  L.J.  Wylie,  A.M.  Coventry and  J.  Seekings (1989) The effect of group 5 chromosomes on the free polar lipids and breadmaking quality of wheat. J. Cereal Sci. 9: 41–51.
•  Nakamura,  H.,  H.  Sasaki,  H.  Hirano and  A.  Yamashita (1990) High molecular weight glutenin subunit of wheat glutenin seed storage protein correlates with its flour quality. Japan. J. Breed. 40: 485–494.
•  Orth,  R.A. and  W.  Bushuk (1972) A comparative study of the proteins of wheats of diverse baking qualities. Cereal Chem. 49: 268–275.
•  Payne,  P.I.,  C.N.  Law and  E.E.  Mudd (1980) Control by homoelogous group 1 chromosomes of the high-molecular-weight subunits of glutenin, a major protein of wheat endosperm. Theor. Appl. Genet. 58: 113–120.
•  Payne,  P.I. (1987) Genetics of wheat storage proteins and the effect of allelic variation on bread-making quality. Ann. Rev. Plant Physiol. 38: 141–153.
•  Pirozi,  M.R.,  B.  Margiotta,  D.  Lafiandra and  F.  MacRitchie (2008) Composition of polymeric proteins and bread-making quality of wheat lines with allelic HMW-GS differing in number of cysteines. J. Cereal Sci. 48: 117–122.
•  Rao,  V.K.,  S.J.  Mulvaney and  J.E.  Dexter (2000) Rheological characterization of long- and short-mixing flours based on stress–relaxation. J. Cereal Sci. 31: 159–171.
•  Roels,  S.P.,  G.  Cleemput,  X.  Vandewalle,  M.  Nys and  J.A.  Delcour (1993) Bread volume potential of variable-quality flours with constant protein level as determined by factors governing mixing time and baking absorption levels. Cereal Chem. 70: 318–323.
•  Santiago,  D.M.,  K.  Matsushita,  T.  Noda,  K.  Tsuboi,  D.  Yamada,  D.  Murayama,  H.  Koaze and  H.  Yamauchi (2015) Effect of purple sweet potato powder substitution and enzymatic treatments on bread making quality. Food Sci. Technol. Res. 21: 159–165.
•  Shewry,  P.R.,  N.G.  Halford,  P.S.  Belton and  A.S.  Tatham (2002) The structure and properties of gluten: an elastic protein from wheat grain. Philos. Trans. R. Soc. Lond., B, Biol. Sci. 357: 133–142.
•  Shewry,  P.R.,  S.M.  Gilbert,  A.W.J.  Savage,  A.S.  Tatham,  Y.-F.  Wan,  P.S.  Belton,  N.  Wellner,  R.  D’Ovidio,  F.  Békés and  N.G.  Halford (2003) Sequence and properties of HMW subunit 1Bx20 from pasta wheat (Triticum durum) which is associated with poor end use properties. Theor. Appl. Genet. 106: 744–750.
•  Takata,  K.,  H.  Yamauchi,  Z.  Nishio and  T.  Kuwabara (2000) Effect of high molecular weight glutenin subunits on bread-making quality using near-isogenic lines. Breed. Sci. 50: 303–308.
•  Takata,  K.,  Z.  Nishio,  W.  Funatsuki,  T.  Kuwabara and  H.  Yamauchi (2003) Difference in combination between Glu-B1 and Glu-D1 alleles in bread-making quality using near-isogenic lines. Food Sci. Technol. Res. 9: 67–72.
•  Wieser,  H. (2007) Chemistry of gluten proteins. Food Microbiol. 24: 115–119.
•  Yamauchi,  H.,  Y.  Ichinose,  K.  Takata,  N.  Iriki and  T.  Kuwabara (2000) Simple estimation of bread-making quality of wheat flour by modified expansion test under reduced pressure. Journal of the Japanese Society for Food Science and Technology 47: 46–49.
•  Yamauchi,  H.,  Z.  Nishio,  K.  Takata,  Y.  Oda,  K.  Yamaki,  N.  Ishida and  H.  Miura (2001) The bread-making quality of a domestic flour blended with an extra strong flour, and staling of the bread made from the blended flour. Food Sci. Technol. Res. 7: 120–125.
•  Yamauchi,  H.,  T.  Noda,  C.  Matsuura-Endo,  Z.  Nishio,  K.  Takata,  T.  Tabiki,  K.  Saito,  Y.  Oda,  W.  Funatsuki and  N.  Iriki (2003) Improving domestic flour for bread making by blending extra strong (ES) flour. Food Preservation Science 29: 211–220.
•  Yamauchi,  H.,  M.  Ito,  Z.  Nishio,  T.  Tabiki,  S.  Kim,  N.  Hashimoto,  T.  Noda,  S.  Takigawa,  C.  Matsuura-Endo,  K.  Takata et al. (2007) Effects of high-molecular-weight glutenin subunits on the texture of yellow alkaline noodles using near-isogenic lines. Food Sci. Technol. Res. 13: 227–234.
•  Yamauchi,  H.,  D.  Yamada,  D.  Murayama,  D.M.  Santiago,  Y.  Orikasa,  H.  Koaze,  Y.  Nakaura,  N.  Inouchi and  T.  Noda (2014) The staling and texture of bread made using the Yudane dough method. Food Sci. Technol. Res. 20: 1071–1078.
•  Zhang,  X.,  H.  Jin,  Y.  Zhang,  D.  Liu,  G.  Li,  X.  Xia,  Z.  He and  A.  Zhang (2012) Composition and functional analysis of low-molecular-weight glutenin alleles with Aroona near-isogenic lines of bread wheat. BMC Plant Biol. 12: 243.