Wheat (Triticum aestivum L. and T. turgidum L. ssp. durum) Kernel Hardness: II. Implications for End-Product Quality and Role of Puroindolines Therein

Authors

  • Anneleen Pauly,

    Corresponding author
    • Laboratory of Food Chemistry and Biochemistry and Leuven Food Science and Nutrition Research Centre (LFoRCe), KU Leuven, B-3001 Leuven, Belgium
    Search for more papers by this author
  • Bram Pareyt,

    1. Laboratory of Food Chemistry and Biochemistry and Leuven Food Science and Nutrition Research Centre (LFoRCe), KU Leuven, B-3001 Leuven, Belgium
    Search for more papers by this author
  • Ellen Fierens,

    1. Laboratory of Food Chemistry and Biochemistry and Leuven Food Science and Nutrition Research Centre (LFoRCe), KU Leuven, B-3001 Leuven, Belgium
    Search for more papers by this author
  • Jan A. Delcour

    1. Laboratory of Food Chemistry and Biochemistry and Leuven Food Science and Nutrition Research Centre (LFoRCe), KU Leuven, B-3001 Leuven, Belgium
    Search for more papers by this author

Direct inquiries to author Pauly (E-mail: anneleen.pauly@biw.kuleuven.be)

Abstract

 Wheat kernel hardness is a major quality characteristic used in classifying wheat cultivars. Differences in endosperm texture among Triticum aestivum L. or between T. aestivum and T. turgidum L. ssp. durum cultivars profoundly affect their milling behavior, the properties of the obtained flour or semolina particles, as well as the quality of products made thereof. It is now widely accepted that the presence, sequence polymorphism, or absence of the basic and cysteine-rich puroindolines a and b are responsible for differences in endosperm texture. These proteins show features in vitro, including foaming and lipid-binding properties, which provide them with a potential impact in the production of wheat-based food products, where they may improve gas cell stabilization or modulate interactions between starch, proteins, and/or lipids. We here summarize the impact of wheat hardness on milling properties and bread, cookie, cake, and pasta quality and discuss the role of puroindolines therein.

Introduction: The Importance of Wheat Hardness for End-Product Quality

Many different food products are made from wheat flour, the particulate material obtained by milling wheat kernels (Delcour and Hoseney 2010). One of the major wheat properties determining end-product quality is its hardness, defined as the force needed to crush the kernels. Common hexaploid wheat (Triticum aestivum L.) endosperm texture ranges from very soft to hard, whereas the tetraploid durum wheat (T. turgidum L. ssp. durum) presents the hardest kernels of all wheat cultivars (cvs.). In Pauly and others (2013a), we focused on wheat hardness and its relation with the presence and functionality of the basic and cysteine-rich proteins puroindoline a (PINA) and b (PINB). As in Pauly and others (2013a), we also here use the North American terminology that distinguishes soft and hard T. aestivum wheat from durum wheat cvs.

Differences in endosperm texture affect the properties and quality of flour and, hence, its preferential use. Flour characteristics such as particle size and damaged starch level determine its suitability for specific end-products (Posner 2000). Another major difference between cvs. indirectly related to hardness is protein quantity and quality. Generally, soft wheat cvs. have been bred to yield flour containing less protein than hard wheats, about 8% to 11% versus 10% to 14% protein, respectively (Delcour and others 2012). Furthermore, hard wheat cvs. have been selected for high water absorption and, hence, for thicker endosperm cell walls. The arabinoxylan (AX) that occurs in wheat endosperm cell walls indeed absorbs large amounts of water. Soft wheat cvs., on the other hand, preferably do not absorb such high levels of water and are thus selected for low water absorption and thin endosperm cell walls containing lower AX levels (Delcour and Hoseney 2010). In general, soft wheat flour is used for producing cake and cookies, hard wheat flour for bread, and durum wheat semolina, namely the purified middlings from T. turgidum which is coarser than T. aestivum flour (see section “Milling of wheat and resulting flour properties”) for pasta (Figure 1), although bread can also be made with soft or durum wheat flour, and cookies and cakes can also be made from hard wheat flour. In noodle production, there are no strict requirements concerning hardness of the used cvs. (Bushuk 1998b). Besides use of flour from cvs. of different hardness, another difference between pasta and noodles is that noodles contain salt(s) in addition to flour and water. North American noodle formulas also contain eggs (Delcour and Hoseney 2010).

Figure 1.

End-product use of wheat cultivars of varying hardness (Bushuk 1998b; Sievert and others 2007; Wrigley 2009).

Some properties of puroindolines (PINs) are at the basis of their potential impact in food processing. These include high surface activity and excellent foam-forming and stabilizing properties, even in the presence of lipids (Clark and others 1994; Dubreil and others 1997; Biswas and others 2001a). In this context, it is worthwhile to mention that intact PINs still occur in baked and stored food products (Capparelli and others 2005), even if part of the endogenously present PINs denature during baking (unpublished results).

This review focuses on the profound impact of wheat hardness on end-product quality and discusses the potential role of PINs in cereal-based end-products. One should also keep in mind that PINA, PINB and its mutant proteins have different structure and lipid-binding properties, which is reviewed in detail in an accompanying paper (Pauly and others 2013a). Logically, this could also result in different impact on end-product quality.

Milling of Wheat and Resulting Flour Properties

Wheat is generally milled to separate the starchy endosperm from the outer layers and germ, thereby reducing the size of the starchy endosperm to obtain flour (or, as stated above, semolina in the case of durum wheat cvs.). Wheat kernel texture is a major determinant of its milling behavior and one of the key properties that need to be known by the miller when maximizing flour yield (Posner 2009). Prior to milling, the kernels are tempered to soften the endosperm and plasticize the bran, which makes the latter less susceptible to fragmentation. Different moisture contents are applied depending on wheat kernel hardness: soft wheat is typically tempered to 14.5% to 15.5%, whereas hard wheat to 15.5% to 17.0% and durum wheat cvs. even to moisture contents above 17.0% (Posner 2000; Manley and others 2011). In the milling process, kernels and kernel fragments are broken, the obtained fragments are further reduced in size and sieved. Wheat hardness profoundly affects all of these operations. For an extensive overview of the different unit operations in milling and the consequences of wheat hardness for the milling process, the interested reader is referred to Posner (2000; 2009). Briefly, the objective of the break rolls is to open the kernels, whereas in the next stage of the milling process, the reduction rolls further reduce the size of the endosperm fragments.

In soft wheat kernels, fracture occurs mainly through the cell contents due to weaker binding of the starch and protein matrix, as discussed in Pauly and others (2013a). In contrast, in hard wheat kernels, the starch granule surface and protein matrix are tightly bound, which causes the fracture to occur mostly at the cell wall (especially in the cells just below the aleurone). In addition, when fracture occurs through the cell contents, due to the stronger starch–protein interaction in hard wheat than in soft wheat, it occurs more often through starch granules in the former case, and results in a higher level of damaged starch (Greer and Stewart 1959; Delcour and Hoseney 2010). Flour from soft wheat and hard wheat typically contains 2.0% to 4.0% and 5.0% to 10.0% damaged starch, respectively (Finney and others 1988; Lin and Czuchajowska 1996). The differences in adherence of protein to starch are also clear in scanning electron microscopy (SEM) images of flour particles (Figure 2) and isolated starch granules (Barlow and others 1973). Kernel hardness is negatively correlated with break flour yield (Rogers and others 1993; Martin and others 2001) and flour particle size is determined by the hardness of its parent wheat (Figure 2). Flour from hard wheat cvs. typically shows a unimodal particle size distribution (with a peak around about 120 μm), whereas flour from soft wheat cvs. displays a rather bimodal particle size distribution (particle sizes of about 25 and 110 μm, with the smaller particles mainly representing free starch granules) (Posner 2009). Particle size affects flour hydration: the finer the particles, the higher the specific surface area per weight unit and, thus, the higher the rate of hydration and water absorption (Bushuk 1998a; Manley and others 2011).

Figure 2.

Particle size distribution (A) of flour from a soft () and hard (---) wheat cultivar. In soft wheat flour, free starch granules occur next to larger flour particles, resulting in a bimodal distribution. The differences between soft and hard wheat flour are also visible in scanning electron microscopy (SEM) images (B; scale bars represent 10 μm), as the soft wheat flour contains more free starch granules and smaller particles. SEM further illustrated that more material adheres to hard wheat starch granules.

Milling the very hard-textured durum wheat into flour inevitably results in excessively high damaged starch levels. To avoid this, durum wheat is usually milled into semolina that is coarser than T. aestivum flour (in the ranges of 180 to 600 μm, 180 to 475 μm, and <350 μm for coarse, middle, or fine semolina, respectively) (Posner 2000).

Since damaged starch and AX absorb high levels of water, wheat hardness has a major impact on the flour water absorption and, hence, the optimal amount of water needed for preparing dough of desired consistency (Greer and Stewart 1959; Bettge and Morris 2000). While undamaged starch granules absorb only about 30% of their weight in water, damaged starch can take up about its own weight (Manley and others 2011). Proteins and AX absorb even more water: about 2 and up to 10 times their own weight, respectively (Bushuk 1998a). Furthermore, since wheat hardness affects flour yield, it evidently also determines flour composition. Protein and AX levels increase with flour extraction rate (Orth and Mander 1975; Dornez and others 2006), which, in turn, increases flour water absorption (Orth and Mander 1975).

Finally, flour ash content, an indicator of flour contamination with bran, is related to kernel hardness. However, flour ash content also depends on kernel size and bran thickness (Posner 2009). Table 1 lists the flour properties related to wheat hardness which are desired in the production of bread, cookies, cakes, and pasta, respectively. The impact of these flour properties on product quality is discussed in the following sections.

Table 1. Overview of flour properties associated with wheat hardness and preferential use
 BreadCookiesCakePasta
  1. a

    Dry matter basis;

  2. b

    Delcour and Hoseney (2010);

  3. c

    Goesaert and others (2005);

  4. d

    Wade (1988);

  5. e

    Miller and others (1967);

  6. f

    de Noni and Pagani (2010);

  7. g

    Sievert and others (2007);

  8. h

    Manley and others (2011);

  9. i

    Kuenzli and others (2008);

  10. j

    Pareyt and Delcour (2008);

  11. k

    Yamazaki and Kissell (1978);

  12. l

    Lai and Lin (2006);

  13. m

    Feillet (1988).

Wheat hardnessHardbSoftbSoftbVery hardb
 (T. aestivum)(T. aestivum)(T. aestivum)(T. turgidum L. ssp. durum)
Flour properties    
Damaged starch (%, dma)ca. 8c2–4b, d<5eca. 5f
Particle size (μm)ca. 75 (average)gca. 50 (average)hca. 50 (average)h125–630i
Water absorptionHighbLowjHighe, kHighb
Protein level (%, dma)13–15l<10d, j<10b, j>13m

Since PINs form the molecular basis of wheat hardness (Morris 2002; Pauly and others 2013a), they indirectly impact the above-described milling process and conditions as well as the properties of the obtained flour. Several studies have investigated the influence of Pin mutations on these properties. Hard wheat cvs. containing mutated PINB proteins (Pina-D1a/Pinb-D1b up to Pinb-D1g) are generally slightly softer and, hence, produce more break flour than cvs. lacking PINA (Pina-D1b/Pinb-D1a). In line with the above, flour from cvs. with mutated PINB proteins has less damaged starch, smaller particles, lower ash levels, and less water absorption than cvs. having no PINA (Rogers and others 1993; Martin and others 2001; Ma and others 2009; Edwards and others 2010). Furthermore, transgenic cvs. overexpressing PINs yield flour of very low damaged starch and ash levels, low water absorption, and small average particle size (Hogg and others 2005; Martin and others 2007; Wanjugi and others 2007).

Bread

Bread making starts with mixing flour, water, yeast (Saccharomyces cerevisiae), salt, and sometimes other (minor) ingredients into dough, which then undergoes fermentation, proofing, and baking. The quality of flour, the main ingredient in the bread recipe, largely affects bread quality. Next to a large loaf volume, consumers appreciate a fine and homogeneous crumb structure and a pleasant mouthfeel (namely “soft”).

Importance of wheat hardness for bread quality

Bread is usually produced from flour from hard wheat because of its relatively high protein level of good quality. Thus, flour used in bread making contains relatively high damaged starch levels and has large particles. With regard to the former, a certain degree of damaged starch can be useful in bread making, especially in recipes without added sugar. Damaged starch is susceptible to amylolytic breakdown and provides yeast with sufficient fermentable sugars to maintain fermentation. However, too high damaged starch levels and/or an excess of amylase activity yield sticky dough (Delcour and Hoseney 2010), especially in processes requiring long fermentation during which α-amylase has a longer time to act (Carson and Edwards 2009). Barrera and others (2007) further found a negative correlation between damaged starch levels and bread volume which they explained in terms of competition for water between damaged starch and gluten proteins, preventing optimal gluten development and resulting in a loss of gas retention capacity. This resulted in denser bread loaves, which, in turn, affected crumb hardness. Also, damaged starch affects dough rheology by its high water-binding capacity (see section “Milling of wheat and resulting flour properties” ). In bread dough from hard wheat flour, (the sum of undamaged and damaged) starch granules absorbs about 46% of the water, protein about 32%, and AX about 22% (Bushuk 1998a). Flour particle size is of minor importance for bread quality, but evidently impacts flour hydration rate and water-holding capacity. Dough mixing hydrates flour particles, rubs them against each other, and completely disintegrates them so that no intact particles are present in optimally mixed dough (Hoseney 1985; Sievert and others 2007).

However, the most important reason for using hard wheat flour in bread making is its higher protein level and better gluten quality than those of soft wheat flour. Gluten proteins form a continuous, visco-elastic network upon hydration and mixing, which retains the gas produced by yeast during fermentation and provides structural support for the expanding gas cells during baking. Both gluten quantity and quality greatly influence dough rheology, water absorption, mixing time and tolerance, bread volume, and crumb structure (Veraverbeke and Delcour 2002; Delcour and others 2012). An optimal gliadin (which contributes viscosity) -to-glutenin (which contributes elasticity) ratio is essential for excellent bread quality. Insufficiently elastic gluten cannot efficiently retain carbon dioxide produced by the yeast, while too stiff gluten impedes gas cell expansion (Hamer and others 2009; Delcour and others 2012).

Impact of PINs on bread quality

Gas cell stabilization in bread dough

Bread crumb structure depends on incorporation of air during dough mixing. The number and size distribution of the incorporated discrete gas cells mainly depend on the properties of the gluten network (see “Importance of wheat hardness for bread quality” section). Further stages of bread making alter the gas cell distribution and, hence, contribute to crumb structure. Bread dough is a foam structure that is intrinsically unstable due to gas cell coalescence and disproportionation (Gan and others 1995; Mills and others 2003; Sroan and others 2009; Sroan and MacRitchie 2009). Initially, it was thought that the gluten-starch matrix encloses the gas cells during fermentation and early baking stages and that gas cell rupture during baking is due to a strong viscosity increase due to starch gelatinization (Singh and Bhattacharya 2005). However, Gan and others (1990), using SEM, found discontinuities in the gluten-starch matrix surrounding the gas cells appearing after 15 min of fermentation. Several studies now support the hypothesis that a liquid film surrounding the gas cells is responsible for their stabilization when discontinuities appear in the gluten-starch matrix (Figure 3) (Gan and others 1995; Sroan and others 2009; Sroan and MacRitchie 2009). Surface-active molecules such as proteins and lipids can help maintain the integrity of such liquid film surrounding the gas cells (Figure 3). Surface-active proteins can form a two-dimensional continuous film at the air–water interface through intermolecular interactions with neighboring proteins, which gives structural support to the gas cells (Gan and others 1995; Mine 1995). In this respect, several authors (Gan and others 1995; Mills and others 2003; Sahi 2003) have suggested a role for PINs in gas cell stabilization. Indeed, PINs are highly surface-active, small proteins, which favors fast migration (Mills and others 2003), and much in vitro research on their structure and organization at the air/water interface has been conducted. Both PINA and PINB are distributed easily and efficiently at the air–water interface, where they form stable monolayers, even at low concentrations (1.0 mg/mL) (Biswas and others 2001b). PINA has a slightly higher surface affinity than PINB, which is likely due to differences in their tryptophan-rich domains (Bottier and others 2008). Biswas and others (2001a) showed that the induction time adsorbing PINs at air–water interfaces is longer for PINA than for PINB, as confirmed later by Clifton and others (2007). Based on the relationship between induction time and conformational stability of globular proteins, it is assumed that the former is an indication for the time taken by protein to unfold at the interface. The longer induction time for PINA thus implies that it is less flexible and more stable than PINB (Biswas and others 2001b). Still, despite the presence of 5 disulfide bridges (structural characteristics of PINs are reviewed by Pauly and others 2013a), both PINs thus show a certain degree of unfolding. A changed orientation of PINs’ α-helices has been observed at the air–water interface (Bottier and others 2008). Likely, PINs form viscoelastic films of aggregated proteins at the air–water interface (Dubreil and others 2003). Also, PINs spread more easily at the air–water interface at higher ionic strength, which has been explained either in terms of salting-out phenomena or in terms of less intramolecular electrostatic interaction within the proteins, possibly inducing unfolding (Biswas and others 2001b). Furthermore, the amino acid residues of PINA and PINB are more closely packed at the air–water interface at lower pH values due to reduced repulsive interactions (Biswas and others 2001b). Also, PINs are more flexible at low pH (Biswas and others 2001a), possibly due to a changed protein conformation (Le Bihan and others 1996). This may suggest that, in bread dough, in which sodium chloride and yeast (which lowers the pH) are essential ingredients (Delcour and Hoseney 2010), PINs improve gas cell stabilization.

Figure 3.

Transformation of bread dough from foam to sponge-like structure, explaining the “liquid film” hypothesis (Gan and others 1990; Gan and others 1995; Sroan and MacRitchie 2009). After mixing, dough contains discrete gas cells surrounded by a liquid film and embedded in the gluten–starch matrix (A). The gluten–starch matrix fails to enclose the gas cells during later stages of fermentation and early stages of baking (B). At this point, the liquid film takes over gas cell stabilization, due to the presence of surface-active proteins and lipid at the air–water interface. Arabinoxylan (AX) can act by stabilizing the interfacial layer or by increasing the viscosity of the aqueous phase (Courtin and Delcour 2002; Mills and others 2003). During baking, the gas cells further expand until the liquid film is no longer able to surround the gas cells resulting in cell opening (the bread is gas-continuous) and loss of gas retention (C).

PIN-lipid interactions at the air–water interface

As already mentioned, proteins and lipids coexist at the air–water interface. Polar lipids are also highly surface-active, and most of them even more so than proteins. Whereas surface-active proteins stabilize air–water interfaces by forming a viscoelastic film, lipids act by the Gibbs–Marangoni mechanism (Mills and others 2003) which itself relies on the stabilization of the interface by a highly fluid layer of lipids. Whenever deformation of the layer results in a decreased (local) lipid concentration, the lipid molecules migrate to the depleted area to restore the concentration gradient. Generally, mixed protein–lipid interfaces are inherently unstable since proteins and lipids compete for the air–water interface, thereby weakening each other's ability to stabilize it (Gan and others 1995; Marion and others 1998; Mills and others 2003; Sroan and MacRitchie 2009). The presence of lipids interferes with the formation of a viscoelastic protein network, whereas proteins hinder the rapid diffusion of lipids to restore concentration gradients in the case of film deformation. However, PINs have a strong lipid-binding capacity (reviewed by Pauly and others 2013a), but whether and to what extent PIN-lipid interactions affect the properties of interfaces and bread dough is not completely understood. It has been shown that the presence of the zwitterionic lysophosphatidylcholine (LPC) enhances the excellent foam stabilization properties of PINA. As the stabilization effect significantly exceeds that expected based on the sum of the impacts of PINA and LPC, this indicates a synergistic action. The PINA/LPC complex formed is likely more surface-active than noncomplexed PINA. Foam stability is maximal at an LPC-to-PIN molar ratio of about 2 to 3 (Wilde and others 1993). In contrast, although PINA and PINB exhibit similar foaming properties, the latter does not show such enhanced foam formation in the presence of LPC (Wilde and others 1993; Husband and others 1994). This observation may well be explained by the different binding strength of PINA and PINB with polar lipids in general and LPC specifically (see Pauly and others 2013a). Dubreil and others (1997) later showed that the lipid-to-PIN molar ratio at which maximum in vitro foam stability is observed, is different for other polar lipid classes and is lower than the values reported by Wilde and others (1993) for LPC. For total extracted polar lipids, namely phospholipids and glycolipids, enhanced foam stability up to a molar ratio of 0.6 for PINA is observed, while foams of PINB are less stable in the presence of total extracted polar lipids. For PINA foams, stability is hardly affected by phospholipids up to a molar ratio of 1.0. At higher molar ratios, foam stability of the mixture is lower than that of PINA alone (Dubreil and others 1997). A similar trend has been observed for PINB foams in the presence of phospholipids. In contrast, glycolipids are detrimental to foam stability irrespective of their molar ratio to either PINA or PINB (Dubreil and others 1997). However, of all flour endogenous lipids, glyco-lipids are the most beneficial to bread making in terms of volume (Selmair and Koehler 2008; 2010). Furthermore, PINs can insert into polar lipid monolayers in vitro (Kooijman and others 1998; Dubreil and others 2003; Biswas and Marion 2006; Clifton and others 2007; Bottier and others 2008). Their insertion behavior depends on the surface pressure (packing) of the monolayer and polar lipid class, and, thus, likely on the type of monolayer formed. Generally, PINs can better penetrate negatively charged monolayers than neutral monolayers (Dubreil and others 2003; Biswas and Marion 2006). Dubreil and others (2003) even showed that PINA aggregates at monolayers of (synthetic) phospholipids, resulting in a very stable lipoprotein film at the air–water interface (PIN-lipid molar ratio of 0.2).

Taken together, first, PINs (especially PINA) can show improved surface-activity and foam stability in the presence of lipids. Second, PINs, by their lipid-binding capacity, can protect foam structures from destabilization by lipids, since mixed protein-lipid air/water interfaces are unstable (depending on lipid-to-protein ratio), thereby allowing other surface-active proteins, such as members of the amylase-trypsin inhibitor family (Salt and others 2005), to stabilize the gas cell surface. Third, PINs can cooperate with (polar) lipids to form a stable lipoprotein layer at the air–water interface. Based on these phenomena, next to the ratio of polar-to-nonpolar lipids and their levels, as reviewed by Pareyt and others (2011), the lipid-to-PIN ratio can be a key determinant for bread crumb structure. However, one should keep in mind that bread is a much more complex system than the above-described models. In flour, PINs are present at the starch granule surface and associated with polar lipids (Feiz and others 2009; Pauly and others 2012). During dough mixing, they are removed from the granule surface and become incorporated in the gluten network, together with polar lipids (Finnie and others 2010; Pauly and others 2012). However, the impact of this transfer is unclear. Furthermore, wheat flour, and, hence, also dough, contain many other constituents which might impact foam stability and gas cell stabilization (such as other proteins and AX; Figure 3).

Bread dough aqueous phase

It is assumed that the surface-active molecules that make up the liquid film around gas cells are present in the dough aqueous phase. This phase, represented by the so-called dough liquor, can be separated by ultracentrifugation, but only when the dough water content exceeds 30% to 35% of dough weight (Gan and others 1995; Salt and others 2006). Primo-Martin and others (2006) detected the presence of proteins (mainly albumins and globulins), nonstarch polar lipids and nonstarch polysaccharides in dough liquor. However, using mass spectrometry, Salt and others (2005) did not detect PINs in dough liquor from dough made from flour from a hard wheat cv. (Pina-D1a/Pinb-D1b), even if a (minor) fraction of the PINs in flour is water-extractable (Rouillé and others 2005b). In contrast, using a very sensitive enzyme-linked immunosorbent assay, we detected low PIN levels (24 μg PINs/mg dough liquor dry matter basis) in dough liquor isolated from either a soft or a hard wheat cv. (Pinb-D1b; glycine to serine change at position 46) (unpublished results). However, it is unclear at present whether dough liquor components are representative for gas cell-stabilizing components in bread dough. Ultracentrifugation might affect the partitioning of wheat flour components into dough liquor. Lipids, for example, are likely not recovered in dough liquor due to high-speed centrifugation and their low extractability in water. Since PINs are associated with lipids (Dubreil and others 2002), this can also explain their absence or low recovery levels in dough liquor.

PINs become incorporated in the gluten matrix, likely together with polar lipids (Dubreil and others 2002; Finnie and others 2010; Pauly and others 2012). When PINA is added to wheat flour dough, the gas cells are lined with lipids and other surface-active components (mainly proteins), but not PINA, which itself is associated with lipids in the gluten network (Dubreil and others 2002). Li and others (2004) showed that a fraction of the flour polar lipids is present at the gas cell surface, and another fraction in the gluten-starch matrix in association with gliadin proteins. According to McCann and others (2009), phospholipids are likely associated with gliadin proteins, while glycolipids preferentially interact with glutenin. However, the relevance of this study for bread making should be questioned as the presence of lipids in gliadin or glutenin fractions could also be related to their partitioning in the used organic solvents. Dubreil and others (2002) further demonstrated that when PINs are added to flour, less lipids are located at the gas cell surface and lipid vesicles in the gluten–starch matrix are smaller, likely indicating that the matrix contains more, but smaller, lipid vesicles. These results may suggest that the major impact of PINs on bread quality is due to their strong lipid-binding would capacity, which prevent lipids from adsorbing to the air–water interface. In contrast, in dough mixed from defatted flour PINA occurs around gas cells and in the gluten–starch matrix (Dubreil and others 2002). However, it is not clear whether adsorption of PINs at the interface is possible only because they are no longer bound to polar lipids in the gluten–starch matrix, or whether PINs can only adsorb when the interface is free of lipids. It should be noted that both Dubreil and others (2002) and Li and others (2004) used exogenous (added) PINA or lipids to reveal the locations of these components. It is not clear whether these components behave in the same way as PINA and lipids naturally present in flour, and, hence, whether these studies are representative for real bread dough systems. Again, Salt and others (2005), using mass spectrometry, did not identify PINs in dough liquor isolated from defatted flour.

Impact of PINs on dough rheology

Since a portion of PINs become incorporated in the gluten matrix during dough mixing (Dubreil and others 2002; Finnie and others 2010; Pauly and others 2012), they might impact dough rheology. Addition of exo-genous PINs in levels as high as twice their natural level in flour altered dough rheological behavior, as shown by Alveograph studies. Resistance to deformation increases upon addition of PINs, suggesting an increase in protein–protein interactions, whether or not mediated by polar lipids. Opposite effects on dough strength and extensibility have been observed for flour of good or poor bread making quality (Dubreil and others 1998). Addition of PINs to good-quality or poor-quality flour increases or decreases dough strength and extensibility, respectively (Dubreil and others 1998). According to Rouillé and others (2005c), addition of PINs moderately increases strain-hardening behavior, while that of PINs to defatted flour increases strain-hardening behavior. The stability of the gluten–starch matrix, which is responsible for stabilizing gas cells against disproportionation and coalescence in the early fermentation stages, depends on its tendency to strain-harden (Sroan and others 2009). The extent of strain-hardening is indeed positively related to fine bread crumb structure (Rouillé and others 2005c; van Vliet 2008). As already mentioned, when discontinuities appear in the gluten–starch matrix, the liquid film takes over the stabilizing role of the matrix. Thus, PINs may impact bread quality by affecting both the primary (gluten–starch matrix) and secondary (liquid film) mechanism of gas cell stabilization. However, one should also note that in the above-mentioned bread making studies the added PINs had been purified with the nonionic detergent Triton X-114, which is very difficult to remove from protein samples. It cannot be excluded that the presence of trace amounts of this strong detergent impacts the outcome of rheological experiments.

Bread baking trials with reconstituted flour samples

Adding purified PINs to wheat flour yields fermented dough with fine and uniform gas cell distribution, indicating that PINs indeed stabilize gas cells during fermentation. However, an impact has been seen neither on dough volume nor on gas cell growth rate, as measured with magnetic resonance imaging (Rouillé and others 2005a). Baking the fine and homogeneous dough enriched in PINs yields bread with small and fine gas cells, a thin gluten–starch matrix, but an unchanged bread volume. The same effects have been observed using defatted flour (Dubreil and others 1998; Rouillé and others 2005b), indicating that PIN-lipid interactions play a major role in bread quality. Furthermore, addition of PINs to defatted flour yields bread with an even finer crumb structure (Rouillé and others 2005c). This may be explained by the fact that, when lipids are removed, no competition between PINs and lipids takes place at the air–water interface and a stable protein film is formed (see section “PIN-lipid interactions at the air-water interface”). However, again, these reconstitution experiments were performed with purified (exogenously added) PINs, and it is not clear whether they behave in the same way as endogenously present PINs. Possibly, purified PINs might not be associated with endogenous lipids and, therefore, differently impact bread quality.

Yeast activity

PINs display in vitro antimicrobial activity against several bacteria and fungi, as discussed by Pauly and others (2013a), and can interact with the plasma membrane of Saccharomyces cerevisiae (Evrard and others 2008). Minimum inhibition concentrations have, to the best of our knowledge, not been reported, but it must not be excluded that yeast activity during fermentation can partially be disturbed by PINs.

Cookies

Flour, sugar, and fat are the 3 major ingredients of cookies (also called biscuits in some parts of the world). Water is added during cookie dough mixing. Baked cookies have a low moisture content (typically 1.0% to 5.0%) (Pareyt and Delcour 2008). In addition to its flavor and color, cookie quality is mainly determined by its dimensions (both height and diameter), bite (and thus cookie structure and texture), and surface-cracking pattern. Cookies are generally classified based on the type of dough they are made of (Wade 1988). Short dough contains relatively high levels of sugar and fat (up to 30% and 25%, respectively), but a low water content (about 15%). Hard dough has relatively low levels of sugar and fat (up to 14% and 12%, respectively) and a higher water content (up to 30%) (Wade 1988; Manley 2011). Because of their formulation, a gluten network is formed during hard dough but not during soft dough mixing.

Importance of wheat hardness for cookie quality

Superior cookies and biscuits are prepared with soft wheat flour (Wade 1988; Pareyt and Delcour 2008; Pauly and others 2013b). Cookies made from such flour have better appearance and eating quality (Wade 1988) and a more tender bite than those made from hard wheat flour (Pareyt and Delcour 2008). During baking, both spread rate and expansion time are larger for dough from soft wheat flour and, hence, yield cookies with a larger diameter (Abboud and others 1985; Miller and Hoseney 1997). This has been explained in terms of the lower levels of protein, damaged starch, and AX in soft wheat flour. As already mentioned (see “Milling of wheat and resulting flour properties” section), high levels of these flour constituents can absorb relatively high levels of water which impairs dissolution of sucrose and, that way, increases viscosity (Pareyt and Delcour 2008; Pareyt and others 2010). Low-protein flours are also used to obtain soft-textured cookies (Fustier and others 2008; Pareyt and others 2008). Furthermore, the use of soft wheat flour (with smaller particles) for hard dough yields denser cookies with less gluten development during baking than does the use of hard wheat flour (with larger particles). In contrast, the use of soft wheat flour for short dough results in cookies with lower density, more gluten development during baking and, hence, less spread in the oven than the use of hard wheat flour (Manley and others 2011). Igrejas and others (2002) found damaged starch to be the most important parameter predicting semisweet biscuit (hard dough) quality. For these systems, dough viscosity, which is mainly determined by damaged starch levels, affects cookie dimensions and textural and microstructural properties. Cookies from flour from hard wheat cvs. are harder (due to lower porosity and a stronger matrix) than those made from soft wheat flour (Pauly and others 2013b).

Impact of PINs on cookie quality

Little attention has been paid to the role of PINs in cookie systems, although relatively high levels of these proteins are present in soft wheat flour (about 0.1% dm) (Dubreil and others 1998; Turnbull and others 2000). Rogers and others (1993), using near-isogenic wheat lines, found that the presence of PINs predicts whether a certain cv. is suitable for producing sugar-snap cookies. However, it is unclear from their study whether PINs induce differences in baking quality or simply are markers for flour quality.

As mentioned in the accompanying paper by Pauly and others (2013a), the role of PINs in wheat hardness has been explained in terms of modulating the interaction between starch granules and the gluten protein matrix in the wheat kernel. These proteins may exert a similar effect in some types of cookies. In cookies prepared from short dough, the starch does not gelatinize during baking and gluten development is minimal or even absent. It seems reasonable to assume that the gluten-starch interaction initially present in the wheat kernel is at least partially maintained during the process. This way, the presence of wild-type PINs could result in a softer texture and, hence, a softer bite. Malouf and others (1992) demonstrated that tablets from dried and ground dough made from model flour reconstituted from starch, gluten, and water-extractable fractions from hard and soft wheat flours, showed low tensile strength when the starch fraction originated from soft wheat, regardless of the source of the other 2 fractions. This indicated that the proteins associated with soft wheat starch granules (including PINs) highly affect wheat endosperm texture (Malouf and others 1992), and also that after dough mixing at least part of their influence is maintained. Of further importance is that Ryan and Brewer (2006) found sugar-snap cookies prepared with deproteinated starch granules to have harder texture than cookies made with native starch. This confirms that starch granule-associated proteins can mediate the interaction between starch and proteins in sugar-snap cookies. However, it is unclear whether the protein removal procedure that was used (including 0.1% dithiothreitol and 0.1% acetic acid) also removes PINs from the starch granule surface, since PINs are thought to be associated with polar lipids (Pauly and others 2013a). Additionally, since there is a tight relationship between proteins and lipids at the starch granule surface (Debet and Gidley 2006; Pauly and others 2012), it seems logical that also at least part of the lipids have been removed from the starch surface, and, hence, that these lipids could also affect cookie hardness. Papantoniou and others (2004) indeed showed that sugar-snap cookies made from defatted flour have harder texture than cookies made from native flour. In “hard dough” cookie systems (such as semisweet biscuits), gluten is partially developed, resulting in partial removal of PINs from the starch granule surface (Pauly and others 2012). However, gluten does not develop to the extent it does in bread dough due to higher fat and sugar and lower water levels in biscuit dough, so that PINs may still partially exert their modulating effect on starch–protein interactions.

Cakes

The term cake is used to refer to a wide range of sweet baked products generally made from flour, sugar, eggs, and either fat or oil (Wilderjans and others 2013). High-ratio cakes contain more sugar than flour, while low-ratio cakes contain less sugar than flour or equal levels of both (Wilderjans and others 2013). During cake making, batter with low viscosity is formed rather than a dough. Cakes have a moisture content between 18% and 28%, which is lower than that of bread, but higher than that of cookies. Depending on the formula, most of the starch gelatinizes during baking (Hui and others 2006; Delcour and others 2012). Volume and crumb quality (tender and crumbly), determined by its texture and structure, are the most important cake quality parameters.

Importance of wheat hardness for cake quality

Although relatively less flour is used in cake making than in bread or cookie making, it profoundly affects quality. Cake volume is negatively correlated with wheat hardness (Park and Chang 2007). Soft wheat flour is indeed preferred for high-quality cakes, mainly because of its small particle size (Yamazaki and Kissell 1978; Gaines 1985). A smaller flour particle size with concomitant larger specific surface area increases flour water holding capacity and, hence, batter viscosity, which is important for restricting migration and coalescence of gas cells. Proteins, together with damaged starch and AX, profoundly impact batter viscosity by their high water-absorbing capacity (see “Milling of wheat and resulting flour properties” section). Furthermore, batter prepared from flour with larger particles incorporates less air during mixing than batter from flour with smaller particles (Duyvejonck 2012). Therefore, flour particle size is sometimes reduced by pin milling. However, flour particles that are too small yield poor-quality cakes, as high damaged starch levels (above 5%) counteract the positive impact of the reduced particle size (Miller and others 1967; Yamazaki and Kissell 1978). Besides its fine granulation, flour from soft wheat cvs. also has a low protein level (see “Introduction: The importance of wheat hardness for end-product quality” section). In contrast to bread dough making, during cake batter mixing gluten proteins do not form a fully developed network due to the shorter mixing time and higher sugar, fat, and moisture levels in the latter (Donelson and Wilson 1960; Yamazaki and Kissell 1978). Although the gluten proteins are less concentrated than in bread dough, they contribute to the cake structure (Donelson and Wilson 1960; Wilderjans and others 2008). With increasing protein level, cake volume first increases to a maximum and then decreases (Donelson and Wilson 1960; Wilderjans and others 2008). Flour of too high protein content yields cake with an undesired elastic, rather than crumbly, desirable texture.

Impact of PINs on cake quality

As for cookies, cakes are generally produced from soft wheat flour and, hence, contain relatively high PIN levels. However, due to the lower flour content of cake batter than that of cookie dough, cakes contain a lower concentration of PINs than cookies. During mixing, air is incorporated in the aqueous or fat phase of the batter, depending on the mixing method. In multistage mixing, fat crystals melt during the early baking phase, and air cells trapped in the fat migrate to the aqueous phase (Conforti 2006). When air cells are present in the aqueous phase, they are stabilized by egg white proteins by conformational rearrangements and subsequent film formation around the cells (Mine 1995). Dubreil and others (1997; 1998) showed that PINs have even better foam-forming and stabilizing properties than egg white proteins. Moreover, foams made from PIN solutions are less susceptible to destabilization by polar lipids than those from egg white proteins (Dubreil and others 1997). This would indicate that the high levels of polar lipids found in egg yolk are detrimental to the foaming properties of egg white proteins since a mixed protein–lipid interface is formed (Kiosseoglou and Paraskevopoulou 2006), PINs can be considered as excellent foam-forming and stabilizing proteins in cake making. However, as mentioned earlier (see “Importance of wheat hardness for cake quality” section), cakes are typically made from a batter without extensive gluten network formation. Finnie and others (2010) and Pauly and others (2012) showed that, in batter systems, PINs are located at the starch granule surface together with polar lipids. This may very well also be the case in cakes. As in the case of bread (see “Impact of PINs on bread quality” section), it is not clear how PIN-lipid interactions play a role in gas cell stabilization and end-product quality.

Pasta

The term “pasta” is traditionally used to describe products made from semolina of durum wheat (T. turgidum L. ssp. durum) cvs. to which low amounts of water are added to obtain crumbly dough, which is then cold extruded and/or dried. Sometimes, other ingredients such as eggs are included in the formula (Delcour and Hoseney 2010). Since durum wheat cvs. do not possess the D-genome, they do not contain PINs (Morris 2002). Good-quality pasta is firm and resilient after cooking (“al dente” bite), with low surface stickiness, and little if any cooking losses. Other quality characteristics are appearance (including color, absence of dark specks, and surface texture), aroma, and taste (Landi 1995; Troccoli and others 2000; Bruneel and others 2010).

Importance of wheat hardness for pasta quality

Durum wheat is the preferred raw material for pasta. In some countries, it is not even allowed to use T. aestivum flour for pasta production. Pasta produced from milled T. aestivum is not as resistant to overcooking as pasta produced from durum wheat semolina (Cubadda 1989). The high levels of carotenoids in durum wheat give pasta its desired yellow color (Troccoli and others 2000). Additionally, durum wheat cvs. generally have higher protein content than T. aestivum cvs. (Feillet 1988), which is the primary factor associated with superior pasta cooking quality (Feillet and Dexter 1996), although durum wheat proteins generally present poorer bread making quality due to their lower glutenin-to-gliadin ratio than found in T. aestivum cvs. (Feillet 1988). However, there is little literature that clearly describes the impact of protein quality on pasta quality. When drying occurs at high temperatures, protein content is the most important characteristic, while protein quality is of minor importance (Novaro and others 1993; Trocc-oli and others 2000). High-temperature-drying (above 60 °C) induces aggregation of gluten proteins before cooking, which reduces the need for forming a strong gluten network during cooking (D'Egidio and others 1990; Delcour and others 2012). In contrast, protein level and quality have equal importance when pasta is dried at low temperature (Troccoli and others 2000). Weak nonelastic gluten is detrimental for pasta cooking quality, but it is not clear whether high-quality gluten is essential in pasta making. An excessively strong protein network cannot withstand starch swelling during pasta cooking, resulting in cooking losses and sticky pasta after cooking (Bruneel and others 2010). For pasta making, mixing is performed at lower moisture level (about 30%) and within a shorter time than for bread making, resulting in a very crumbly dough with partially developed gluten network and higher importance of other semolina components (Feillet and Dexter 1996; de Noni and Pagani 2010).

As already mentioned, durum wheat cvs. are generally milled into semolina rather than flour to avoid high damaged starch levels (see “Milling of wheat and resulting flour properties” section). Mechanical damage during mixing and extruding increases damaged starch levels during pasta processing (Lintas and D'Appolonia 1973; Vansteelandt and Delcour 1998). Therefore, it is important to limit the damaged starch level in the raw material. High damaged starch levels result in sticky dough and pasta with increased surface stickiness (Grant and others 1993), especially when amylases are present, and induce starch solubilization (de Noni and Pagani 2010). In addition, semolina with a fine and regular particle size (125 to 350 μm) is preferred in pasta production to obtain a uniform water distribution among the semolina particles during processing. In the case of nonuniform water distribution, undesirable white spots are found on the pasta surface (Banasik 1981; Troccoli and others 2000; Posner 2009; de Noni and Pagani 2010). However, there is an optimum between decreasing semolina particle size and, at the same time, increasing damaged starch levels when aiming for high-quality pasta.

Recently, durum wheat has been rendered soft textured by introducing the genes encoding PINs (namely Pina-D1 and Pinb-D1) by nontransgenic means. Gazza and others (2008) were the first to report successful transfer of PIN genes into durum wheat, with an accompanying 60% reduction in single kernel characterization system (SKCS) hardness value. Morris and others (2011) developed stable soft durum lines through ph1b-mediated homoeologous recombination which, in contrast to the former, can be used in future crosses with durum parents. The obtained soft durum line had SKCS hardness values and milling behavior similar to those of T. aestivum soft wheat cvs. (Morris and others 2009; Morris and others 2011). As expected, such soft durum lines significantly decreased water absorption and altered dough rheology, resulting from lower damaged starch levels and smaller particle size (Casper and others 2011; Gazza and others 2011). The decreased water absorption of the soft durum lines reduced the amount of water that must be removed in the pasta drying step, and thus decreased energy inputs and drying times (Casper and others 2011). Additionally, most modern pasta plants prefer small-particle-sized flour or semolina (faster dough hydration and uniform distribution of water among the semolina particles) (Troccoli and others 2000; Posner 2009), which is easily achievable with soft durum lines. Gazza and others (2011) further showed that pasta from soft durum lines has the same bright yellow color and cooking quality (firmness and stickiness) as pasta products from hard durum lines, while Casper and others (2011) found soft durum pasta products to be slightly more cook-tolerant and firmer than products from hard durum lines. This indicates that modulating the durum endosperm texture does not impair its pasta making potential and may even improve its dough mixing performance (Casper and others 2011; Gazza and others 2011).

Impact of PINs on pasta quality

Durum wheat cvs. do not contain PINs due to the lack of the D-genome (Morris 2002). However, with the recent development of soft durum lines, the effect of the presence of PINs in the raw material for pasta making can be evaluated. Gazza and others (2011) described comparable cooking quality for pasta made from soft durum lines and pasta made from traditional durum lines, suggesting that the impact of PINs during pasta cooking is rather low. Superior pasta cooking quality is obtained when the protein network withstands starch swelling during cooking when a competition exists between protein polymerization and starch swelling (Delcour and others 2000; Bruneel and others 2010). Insufficient protein polymerization during cooking results in a discontinuous network (Resmini and Pagani 1983), while in the case of excessive cross-linking, the proteins lack resilience to cope with starch swelling during cooking (Bruneel and others 2010). PINs together with polar lipids can modulate the interaction between starch and gluten, as they weaken the starch–protein interaction during wheat desiccation and maturation (Lillemo and Morris 2000; Feiz and others 2009). In pasta, the gluten network is only partially developed since mixing is performed at relatively low moisture levels and for short times, resulting in a crumbly dough. This, in light of the data by Pauly and others (2012) would suggest that at least part of the PINs are present at the starch granule surface in raw pasta dough. It is not known how sheeting, extrusion, and drying impact distribution of PINs in pasta dough. Additionally, Delcour and others (2000) showed that starch surface characteristics seem to be of little importance for starch interaction behavior, implying that gluten-starch interactions in raw pasta are mainly due to physical inclusion of starch in the protein network.

Conclusion

Endosperm hardness is a major trait for classifying wheat cvs., with major consequences for the whole supply chain. It is widely accepted that the PINA and PINB proteins, together with polar lipids, determine wheat endosperm texture. The presence of PINs indirectly determines wheat milling behavior, mill settings, the properties of the obtained flour, and end-product quality. In contrast to the impact of hardness-associated flour properties (such as damaged starch and particle size) on end-product quality, which has been well studied and is now mostly understood, the role of PINs for end-product quality is not clear. It has been suggested that they play a role in gas cell stabilization or that they modulate starch-protein interactions. Their lipid-binding capacity and excellent interfacial and foaming properties have been demonstrated in vitro however, and it is unclear how PINs behave in cereal-based end-products. The few studies cited in this paper dealing with the impact of PINs on end-product quality used exogenous (purified) PINs. This may not be representative of endogenous PINs, which are associated with polar lipids in flour and (intermediate) wheat-based products. Investigating the impact of endogenous PINs (namely wild-type versus mutant) is challenging, since the presence and functionality of PINs results in different flour properties, also affecting end-product quality. Much research is still needed to elucidate the role of PINs.

Abbreviations
AX

arabinoxylans

cvs.

cultivars

LPC

lysophosphatidylcholine

PINA

puroindoline a

PINB

puroïndoline b

PINs

puroindolines

SDS-PAGE

sodium dodecyl sulfate polyacrylamide gel electrophoresis

SEM

scanning electron microscopy

SKCS

single kernel characterization system.

Acknowledgments

This work is part of the Methusalem programme “Food for the future” (2007 to 2014). Bram Pareyt gratefully acknowledges the Research Foundation—Flanders (FWO—Vlaanderen, Brussels, Belgium) for a position as postdoctoral researcher. Jan A. Delcour is W.K. Kellogg Chair in Cereal Science and Nutrition at the KU Leuven.

Ancillary