Wheat (Triticum aestivum L.) lipid species distribution in the different stages of straight dough bread making

Highlights

• Wheat flour polar and nonpolar lipid distribution was studied from flour to bread.

• Polar and nonpolar lipids were measured with ESI-MS/MS and HPLC-ELSD, respectively.

• Dough mixing redistributes lipids from free to bound lipid extracts.

• Fermentation lowers glycolipid levels but raises nonpolar- and phospholipid levels.

• Fatty acids replace lysophosphatidylcholine in amylose-lipid inclusion complexes.

Abstract

Although wheat endogenous lipids strongly impact bread quality, knowledge on their detailed distribution throughout the different stages of straight dough bread making is lacking. We here compared the lipid populations in hexane [containing free lipids (FLs)] and water-saturated butanol extracts [containing bound lipids (BLs)] of wheat flour, freshly mixed and fermented doughs, and bread crumb using high-performance liquid-chromatography [for nonpolar lipids, i.e. mainly free fatty acids (FFA) and triacylglycerols] and electrospray ionization tandem mass spectrometry (for polar lipids). Freshly mixed doughs had lower FL and higher BL levels than flour, a phenomenon referred to as lipid-binding. Furthermore, probably due to the disintegration of flour particles, the overall extractability of nonpolar lipids was higher in freshly mixed dough than in flour. Dough fermentation decreased the extractability of glycolipids, but increased that of nonpolar lipids and phospholipids. We hypothesize that these phenomena result from stretching of the gluten network due to gas cell expansion, which leads to the replacement of some lipids associated with gluten proteins by others. Baking increased the extractability of bound lysophospatidylcholine (LPC) levels, but decreased that of free FFA. This is probably due to in situ dissociation of amylose-LPC inclusion complexes and formation of amylose-FFA inclusion complexes during bread baking and cooling, respectively.

The approach and ESI-MS/MS methodology we developed provided valuable insights regarding the distribution of lipids at the different stages of bread making. Hence, it opens perspectives for future efforts to relate differences in lipid composition between wheat cultivars to their bread making quality.

Introduction

Although lipids only make up 1.5 to 2.5% w/w of wheat flour, they strongly impact its bread making quality (Chung, Ohm, Ram, Park, & Howitt, 2009; Hargin & Morrison, 1980; Pareyt, Finnie, Putseys, & Delcour, 2011). Wheat flour lipid structure, classification, and functionality has been a topic of debate over the past century. The first reports on wheat lipids were by Sullivan, Near, and Foley (1936) and Olcott and Mecham (1947). Later, several research groups studied the extraction and separation of wheat lipids (Christie, 1985; Christie, 1986; Christie & Morrison, 1988; Morrison & Coventry, 1985; Prieto, Ebri, & Collar, 1992) and their role in bread making (Chung, Pomeranz, & Finney, 1982; Chung & Tsen, 1975; Fisher, Broughton, Peel, & Bennett, 1964; Hargin & Morrison, 1980; Hoseney, Finney, Pomeranz, & Shogren, 1969; MacRitchie & Gras, 1973; Pomeranz, Chung, & Robinson, 1966). More recently, additional lipid extraction procedures (Hubbard, Downing, Ram, & Chung, 2004; Moreau, Powell, & Singh, 2003), analytical techniques (Finnie, Jeannotte, & Faubion, 2009; Finnie, Jeannotte, Morris, & Faubion, 2010; Finnie, Jeannotte, Morris, Giroux, & Faubion, 2010), or innovative approaches, such as those based on synthetic lipid-like compounds (Selmair & Koehler, 2008; Selmair & Koehler, 2009) or lipases (Gerits, Pareyt, & Delcour, 2014; Gerits, Pareyt, Masure, & Delcour, 2015; Schaffarczyk, Østdal, Matheis, & Koehler, 2016), have allowed further elaborating on wheat lipid structure, classification, and functionality.

Today, wheat lipids are typically classified either as starch lipids, which - as their name implies - occur inside starch granules, or as non-starch lipids (Morrison, 1981). The latter are further subdivided in free lipids (FLs) and bound lipids (BLs) based on their sequential extractability with nonpolar (e.g. hexane) and polar [e.g. water saturated butanol (WSB)] solvents, respectively (Chung et al., 2009; Morrison, 1988; Pareyt et al., 2011). In addition, flour lipids are also classified as either nonpolar or polar. Fig. 1 shows the chemical structures of the most common wheat flour lipid classes. The most abundant flour nonpolar lipids are triacylglycerols (TAGs) and free fatty acids (FFA). Glyceroglycolipids and glycerophospholipids (further referred to as glycolipids and phospholipids, respectively) make up most wheat flour polar lipids. Mono- (MGDG) and digalactosyldiacylglycerols (DGDG) are the main glycolipids, whereas N-acyl phosphatidylethanolamine (NAPE), phosphatidylethanolamine (PE), phosphatidylcholine (PC), and their respective lysoforms lyso-NAPE (NALPE), lyso-PE (LPE), and lyso-PC (LPC) represent the most abundant phospholipid classes (Finnie et al., 2009; Hargin & Morrison, 1980; Pareyt et al., 2011).

Evidently, wheat flour contains a complex mixture of lipids with varying polarities. Some of its lipids play a prominent role in each stage of the bread making process which typically starts by mixing water, flour, yeast, and salt and some nonessential ingredients into viscoelastic dough. During mixing, air is incorporated in the dough (Baker & Mize, 1941; Delcour & Hoseney, 2010). In this process also, gluten proteins interact with one another and form a viscoelastic network while, at the same time, the majority of native flour lipids is redistributed from the surface of starch granules to this gluten network (Chung & Tsen, 1975; Gerits, Pareyt, & Delcour, 2013; Olcott & Mecham, 1947). This phenomenon is known as lipid-binding (Carr, Daniels, & Frazier, 1992; Gerits et al., 2013; Olcott & Mecham, 1947; Ponte, Titcomb, & Cerning, 1964). In the early stages of fermentation, gas cells are embedded in and physically stabilized by the gluten-starch matrix (Gan, Ellis, & Schofield, 1995; Sroan, Bean, & MacRitchie, 2009). However, during late fermentation and early baking, stretching of the gluten network results in discontinuities in the matrix that leave neighboring gas cells only separated by a thin liquid film. From that moment onwards, the stabilization of these gas cells at the air/water interface (Gan et al., 1995; Sroan & MacRitchie, 2009) is taken over by protein and lipid surface-active constituents. During the initial phases of baking, gas cells continue to expand until the liquid films fail to withstand the increase in interfacial area and the gas cells rupture which more or less coincides with the setting of the crumb as a result of starch gelatinization and gluten polymerization. Starch gelatinization provokes (i) migration of water from the gluten to the starch phase which then leads to release of polar lipids from the gluten phase (Eliasson, 1985; Köhler, 2001) and (ii) the in situ dissociation and reformation of amylose-lipid (AM-L) inclusion complexes during bread baking and cooling, respectively (Goderis, Putseys, Gommes, Bosmans, & Delcour, 2014; Kugimiya, Donovan, & Wong, 1980; Putseys, Lamberts, & Delcour, 2010).

It follows from the above that lipids play a decisive role in all stages of bread making. However, because the role of certain lipid classes or even lipid species in the process remains unclear at present, there is a need for a systematic approach for characterizing and analyzing lipids, especially those present in complex matrices (Wenk, 2005) such as wheat flour, dough, or bread crumb. This area of study in general is referred to as lipidomics or lipid profiling (Finnie et al., 2009; Finnie, Jeannotte, Morris, & Faubion, 2010; Finnie, Jeannotte, Morris, Giroux, & Faubion, 2010). Over the last decade, several advanced proteomic techniques [e.g. high-resolution chromatography, mass spectrometry (MS), and nuclear magnetic resonance] have become available (German, Gillies, Smilowitz, Zivkovic, & Watkins, 2007). MS-based lipidomics allows simultaneous identification and quantification of (hundreds of) lipid species in crude lipid extracts (i.e. shotgun lipidomics) [as reviewed by Wenk, 2005, Wenk, 2010 and Dehairs, Derua, Rueda-Rincon, & Swinnen, 2015]. Lipid profiling of wheat whole meal, flour, and starch (Finnie et al., 2009; Finnie, Jeannotte, Morris, & Faubion, 2010; Finnie, Jeannotte, Morris, Giroux, & Faubion, 2010), milling and pearling fractions (González-Thuillier et al., 2015), and flour, dough liquor, and dough liquor foam (Salt et al., 2018) with electrospray ionization tandem MS (ESI-MS/MS) already allowed identification and quantification of different wheat glycolipid and phospholipid classes and the analysis of their distinctive acyl groups. However, detailed knowledge on the distribution of glycolipids and phospholipids during the different phases of bread making is lacking. Furthermore, none of the above studies have included the quantification of NAPE and NALPE, even if they have a unique molecular structure amongst wheat phospholipids because they contain a fatty acid (FA) attached to the sn-3 position.

Recently, a method based on single-run high-performance liquid-chromatography (HPLC) with evaporative light scattering detection (ELSD) was developed at our research group to study wheat lipids (Gerits et al., 2013). Although this method is valuable for studying the role of wheat lipids in bread making (Gerits et al., 2013; Gerits et al., 2014; Gerits et al., 2015), shortcomings include that it is not quantitative and cannot differentiate between individual lipid species within a lipid class. Therefore, we here implemented an MRM-based ESI-MS/MS method for polar lipid profiling, including NAPE and NALPE, thus complementing the HPLC-ELSD lipid analyses, which are still used to assess the distribution of nonpolar lipids. Using this approach, we here set out to study the distribution of polar and nonpolar lipids throughout the entire bread making process by analyzing and comparing lipid extracts from wheat flour, fresh and fermented dough, and bread crumb.