High molecular weight glutenin subunits of wheat : qualitative and quantitative variation in relation to bread-making quality

In view of the poor bread-making quality of the wheat grown in The Netherlands, only a small part of production is used for baking of bread. Therefore quality improvement is a major aim of plant breeding. Unfortunately, breeding for breadmaking quality is hampered by its complexity. The suitability of wheat flour for the manufacture of bread depends on the composition of a large number of kernel components, such as lipids, starch and proteins, and this kernel composition in turn depends on agronomic measures, climate and on the genetic make-up of the varieties. Moreover the suitability of wheat flour for the baking of bread depends on technological conditions, such as those in milling and in the baking process. Another drawback in breeding for improved bread-making quality is the lack of simple and fast techniques to assess bread-making quality. Dough and baking- tests, which take much labour and time are required to determine such characteristics as dough strength and elasticity, loaf volume and crumb structure.Plant breeders in various countries have been aware that the kernel proteins play a major role in determining the bread-making quality. The total <em>amount</em> of kernel proteins, which is largely determined by environment and as such difficult to change by breeding, is positively correlated with the bread-making quality. More useful characteristics for breeding are the genetically determined differences in protein <em>quality,</em> which result in differences in bread-making quality at a given amount of protein, i.e. in differences in the ratio between loaf volume and protein content. Small-scale tests to estimate protein quality (e.g. the SDS sedimentation, the Pelshenke dough-ball or the residue-protein tests) have been used in breeding programmes to predict bread-making quality. However the results of these tests not only depend on the genotype but also reflect environmental conditions. After the development of biochemical techniques for separation of proteins, bread-making quality has been related to the composition of kernel proteins. Because protein composition is genetically determined, it has been proposed as a tool in breeding for bread- making quality.In this thesis the bread-making quality of wheat was studied in relation to differences in type and amount of the various high molecular weight glutenin (HNWg) subunits. These storage proteins are encoded by genes at three homoeologous loci ( <em>Glu-1</em> loci: <em>Glu-A1</em> , <em>Glu-B1</em> and <em>Glu-D1</em> ); at each locus several alleles occur that differ in type or number (0-2) of subunits encoded. At the start of this study, the relation between variation in the HMWg subunit genotype and breadmaking quality has been studied extensively by various research groups. These studies revealed the occurrence of alleles that differ in their contribution to breadmaking quality. In fact, a large part of the variation in bread-making quality was ascribed to variation in the HMWg genotype. Though it was known that variation in type of the HMWg subunits also affects the quality of wheat grown in The Netherlands, it was not known to what extent differences in bread- making quality can be ascribed to differences in the HMWg subunit genotype.Two general aspects for breeding for bread-making quality required further research:1) the difficulties which may occur in the identification of the HMWg subunit alleles2) the variation in amount of the HMWg subunits on which aspect very little information was available.Variation in the amounts of groups of kernel proteins, such as the gliadins and the glutenins, is of prime importance for bread-making quality (Chapter 1). Therefore, it was worthwhile studying whether genetic variation in the <em>amounts</em> of the HMWg subunits could also be used in breeding for bread-making quality. Variation in the amounts of the subunits is probably also important for other applications of wheat in which gluten characteristics are important, such as the manufacture of pastas and pastry. Another reason for studying the HMWg subunits quantitatively is that the cause of the 'differences' in quality of HMWg alleles is not known. Differences in the structure of the subunits (the intrinsic <em>quality</em> of the subunits) have beensuggested, but also differences in the level of gene expression (the <em>quantity</em> of the subunits) could be responsible for the differences in quality between the alleles. Therefore a major aim was to study the variation in amount of individual HMWgsubunits and to elucidate the mechanisms underlying this variation.Chapter 2 shows that variation in the HMWg genotype does have a large effect on bread-making quality of wheat lines grown in The Netherlands. These lines represent a wide genetic background and can be considered to be random lines with respect to bread- making quality. A baking test developed in The Netherlands was used to determine quality. In general, the ranking of the alleles for quality was in agreement with the ranking found in studies in other countries. Because different baking tests and different criteria for quality were used in the various studies, the relationship found can be considered as generally valid. Most important is the variation at the <em>glu-D1</em> locus. The allele encoding the subunits 5+10 is related to a higher bread- making quality than its allelic counterpart encoding 2+12. In the literature, it is generally assumed that the effect of allelic variation at the 3 homoeologous <em>(Glu-1 loci</em> on quality is additive. The so-called scoring systems to screen lines for bread-making quality are based on this assumption. However this study shows that interactions between homoeologous loci do occur: an effect of allelic variation at the <em>Glu-A1</em> and <em>Glu-B1</em> loci was present only in combination with the 'high-quality' <em>glu-D1</em> allele encoding the subunits 5+10. The mechanisms underlying these interactions are not known. At this point, it should be noted that in Dutch wheat varieties, 'poor quality' alleles predominate at each of the three <em>(Glu-1</em> loci (Chapter 7). Therefore it is important to introduce 'high-quality alleles' into Dutch breeding programmes, in particular the <em>glu-D1</em> allele encoding 5+10. About 20% of the variation in the breeding lines was ascribed to differences in HMWg genotype, which is much lower than in some of the other published studies. The remainder of the variation in bread-making quality was ascribed to variation in environmental conditions and to genetic variation in the composition of other kernel components, such as lipids, starch, gliadins and low molecular weight glutenin subunits. Therefore, genetic variation in other kernel components should also be taken into consideration when selecting for bread-making quality, although the HMWg subunit genes can be considered as major genes in determining that quality. Which of these other components is most important in this respect is not known.A large number of HMWg alleles are present in varieties and in species related to bread-wheat. Because the contribution to quality is known for only a small number of these alleles, there could be alleles with an unknown 'quality' in breeding programmes. Unintended introduction of 'poor-quality' alleles can be avoided as far as these alleles can be distinguished from 'high-quality' ones. The HMWg alleles are routinely identified by determining the relative mobility (R _m ) with one-dimensional sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) of the subunits encoded. As shown in Chapter 3, the resolution of this technique may be insufficient for an unambiguous identification because alleles may be present that differ only slightly in the R _m of the subunits encoded. Unambiguous identification by two-dimensional electrophoresis is especially necessary before introducing species related to bread-wheat in breeding programmes.To study variation in the <em>amount</em> of <em></em> the HMWg subunits effectively, one must develop a method for a reliable quantification of <em></em> the subunits present in the kernel. Since one of <em></em> the objectives was to investigate the amounts of <em></em> the subunits encoded by different alleles and because a genotype may possess 'poor-quality' alleles in combination with 'high-quality' alleles, one must use a method that allows the quantification and identification <em>of</em> each subunit. In most methods described in the literature, reversed-phase high-performance liquid chromatography (RP-HPLC) is used to separate the HMWg subunits. However the sample preparation required before RP-HPLC is a potential source of error and may even result in systematic losses of the HMWg subunits, as discussed in Chapter 4. In addition, these methods require electrophoretic analysis to identify the HMWg subunits present in the RP-HPLC peaks. As an alternative, an efficient method was developed and is described in Chapter 4. The HMWg subunits are separated from other kernel proteins - and from each other - by SDS-PAGE. Scanning densitometry is used to determine the amount of <em></em> Coomassie brilliant blue (CBB) adsorbed by the protein after get staining, which is a measure for the amount of protein. The routine method for gel staining resulted in subunit bands with a low staining intensity. So it was essential to introduce a new and more sensitive staining method. Since sample preparation is not required in the method, systematic losses of <em></em> the HMWg subunits can be avoided. Because the identification of the subunits is based on their R _m during SDS-PAGE, the method allows quantification and identification of the subunits in one analysis.As shown in Chapter 5, the amount of individual (HNWg) subunits is subject to environmental variation. An increase in nitrogen fertilizer applied to two varieties (Obelisk and Urban), grown at the same trial field, resulted in an increase in both the total amount of kernel proteins and the amounts of all HMWg subunits. Variation in protein content accounted for 88% and 87% of the variation, respectively, in amount of <em></em> the HMWg subunits. For wheat varieties grown at different sites, there was a weaker correlation between protein content and the total amount <em>of</em> HMWg subunits 'within varieties/between sites'. The proportions of <em></em> variation in amount of <em></em> the subunits accounted for by protein content were 40, 0, 66 and 0% for the varieties used, Citadel, Granada, Okapi and Kraka, respectively. This means that the proportion of these subunits relative to the total amount of kernel proteins, is highly variable as a result of differences in growing conditions (Figure 5.2). The mechanism responsible for this variation is not known. As discussed in Chapter 1, this proportion of the HMWg subunits largely determines bread-making quality of a flour. The environmental effect on the proportion of the HMWg subunits may therefore be useful as an additional criterion for prediction of bread- making quality of flours by millers and bakers, besides traditional criteria such as protein content and variety. Genetic variation in the total amount of the HMWg subunits also plays a role (Chapter 5). This will be discussed later.The environmental effects on the variation in the amount of the HMWg subunits and on the proportion of the subunits relative to the total content of protein hampers detection of genetic variation in amount. However the ratio between the individual subunits of a genotype is only slightly affected by differences in environment. Therefore one can use the amount of a subunit relative to the total amount of the subunits as a measure for the level of gene-expression.The proportion of the HMWg subunits relative to the total amount of proteins can be changed by several breeding strategies. Firstly, it is possible to select for number of subunits because alleles occur at each <em>Glu-I</em> locus that differ in the number of subunits encoded. The number of subunits affects the total amount of the HMWg subunits. Secondly, it is shown in Chapters 5 and 6 that the level of expression of the HMWg subunit genes is affected by variation in their genetic background. For instance, the proportion of the HMWg subunits of two varieties (Citadel and Kraka) with the same HMWg genotype, averaged over six growing sites, differed by about 30% (Chapter 5). This variation offers ways of changing the proportion of the HMWg subunits by breeding. However to achieve this, more information about the nature of this effect of genetic background is required, and techniques for determination of the proportion of the HMWg subunits in breeding programmes must be developed. Thirdly, one can use HMWg alleles that differ only in their level of gene expression, and not in the R _m or the number of subunits encoded ( <em>expression</em><em>alleles</em> ). <em></em> Because the introduction and selection for these alleles is relatively simple, this approach seems more practical than using the effect of variation in genetic background. Expression alleles can also be used to increase the amount of the (HNWg) subunits if the maximum number of subunits is present, or to improve the ratio between the individual HMWg subunits. Selection for expression alleles allows simultaneous selection for type of subunits. It is not yet possible to increase the amount of the HMWg subunits by gene technology (e.g. the introduction of extra copies of the genes).Chapters 6 and 7 study genetic variation in expression of individual HMWg subunit genes. Chapter 6 describes experiments to determine the effect of variation at <em>Glu-I</em> loci on the expression of HMWg genes at homoeologous loci. In theseexperiments, near-isogenic lines that differ in their HMWg genotype, were used and also the progeny of two crosses. The expression of the individual HMWg subunit genes is sometimes, but not always, affected by the number of subunits. Introduction of a <em>glu-D1</em> null-allele resulted in an increase in expression, whereas introduction of a null-allele at the <em>Glu-A1, locus</em> had no effect. Variation in the type of the HMWg subunits had, with one exception, no significant effect on the amount of the subunits encoded by homoeologous genes. Although variation in the genetic background strongly affected expression of HMWg genes, the ratio between the individual HMWg subunits of lines with the same HMWg genotype is only slightly affected by differences in genetic background. An enhancer element, localized proximate to the transcription initiation site of the genes, is thought to be involved in the developmental regulation of the individual HMWg genes. Since variation in the genetic background and in environmental conditions affects the total amount of the HMWg subunits but does not change the ratio between the subunits, expression of the individual HMWg homoeologous genes is probably coordinated by a common regulation mechanism. This mechanism may differ between genetic backgrounds. Coordination of gene expression may involve regulatory proteins, but it may also only be the strength of the gene promoters that is of importance. Ile effect of the genetic background could then be caused by a non-specific mechanism, such as competition for transcriptional factors between the (HNWg) genes and other genes in the genetic background.The relative amounts of the subunits can be used as a measure of the genetically determined level of gene expression because variation in environmental conditions, in genetic background and in (HNWg) genotype has only small effects on the ratio between the subunits. In a set of 38, mostly Dutch, wheat varieties the variation in relative amounts of the HMWg subunits was studied (Chapter 7). The subunits were identified according to their R _m during SDS-PAGE. There were only small differences in the relative amounts of identical subunits. So expression alleles were not detected in this set of varieties. This could be due to a common ancestry of the varieties studied, so one should study the expression of HMWg genes in a set of genotypes of wider genetic origin. However the occurrence of expression alleles may turn out to be generally rare.Non-identical subunit alleles at the <em>Glu-B1 locus</em> of the Dutch varieties differed considerably in the relative amounts of the subunits encoded, as indicated by the relative absorbance of the CBB stained bands. Chapter 8 shows that these differences in CBB adsorption reflect differences in the level of gene expression (i.e. molar amounts of the subunits), and not differences between the subunits in their affinity for CBB nor in their molecular weight. The contribution to breadmaking quality of these <em>Glu-B1</em> alleles, as established in Chapter 2 and in the literature, is positively related to differences in the amounts of the subunits. For the <em>Glu-A1</em> and <em>glu-D1 loci,</em> there was no clear evidence for such a relation. As discussed in Chapter 7, it is still possible that expression alleles occur at the <em>Glu-A1</em> locus, which could explain the conflicting ranking of the alleles found in the literature for quality of the alleles encoding the subunits I and 2*. Because only a few varieties used in Chapter 7 contained these alleles, this could not be investigated for Dutch varieties. The results of this thesis do not unambiguously show differences in the amounts of the allelic <em>glu-D1</em> pairs of subunits 2+12 and 5+10. In Chapter 6, the total amount of the 'high-quality' subunits 5+10 is somewhat higher than the total amount of the 'low-quality' subunits 2+12; in Chapter 7, there is, however, no difference in the total amounts of these subunits. The <em>glu-D1</em> alleles present in the genetic stock used (Dutch varieties) probably do not differ in their level of gene expression. Consequently, the question whether differences in 'quality' between HMWg alleles are caused by differences in the amounts of the subunits encoded or by differences in the intrinsic quality of the proteins remains to be answered. Therefore further research on the amounts of the subunits is required, especially for those encoded by alleles at the <em>Glu-A1</em> and <em>glu-D1 loci.</em> If the relationship found between the amount of the <em>Glu-B1</em> subunits and bread-making quality can be confirmed for other <em>Glu-B1</em> alleles and for the alleles at the other loci, this relationship could be used for prediction of the 'quality' of novel alleles.The following example, based on Chapter 8, illustrates that the HMWg subunits that are of crucial importance for the bread-making quality only constitute a minor fraction of the flour proteins (on a weight basis). An estimate of the proportion of the <em>Glu-B1</em> subunits (relative to the total amount of kernel proteins) of a genotype containing the 'high-quality' allele encoding 7+9 is 3% (Table 8.4). Accordingly, I kg of a flour (protein content 12%) of this genotype contains 3.6 g of these 'highquality' subunits. The amount of the 'low-quality' subunits 6+8 <em>Glu-B1</em> allele) is about 80% of the amount of 7+9 (Table 7.4). So I kg of a flour milled from a genotype possessing this allele contains 2.9 g of these subunits. In the research presented in Chapter 2, the differences in loaf volume of lines that possess these <em>Glu-B1</em> alleles is about 20%. Whether or not the differences in the amount of the subunits are responsible for the differences in quality of these alleles, this example illustrates that differences in a minor flour fraction can have large effects on the bread-making quality.In conclusion, not only qualitative variation but notably also quantitative variation in the composition of the HMWg subunits is important for bread-making quality. This quantitative variation in the HMWg composition probably also affects the suitability of wheat for other purposes, e.g. in the pasta, pastry, biscuit and starchgluten industry. Because genetic variation in the amount of the HMWg subunits occurs, it is possible to improve the proportion of the HMWg subunits by breeding. For example, if the average gluten strength is too low for good quality, a breeder can increase the amount of the subunits. Improvement by breeding of the quantitative (HNWg) composition can go together with breeding for type of the subunits. The results also show that the proportion of the HMWg subunits relative to the total amount of flour proteins is subject to environmental variation, which is important for improvement of the quality of wheat for the various applications, but it cannot be used fully because simple and fast tests are lacking for estimation of the HMWg subunits quantitatively.Before the qualitative and quantitative variation can be exploited optimally in improving bread-making quality of wheat, additional research will be required. The proportion of the HMWg subunits required for good quality will depend on the amount and the composition of other kernel proteins, such as gliadins and low molecular weight glutenin subunits and probably also on the type of HMWg subunit. The optimum ratio between groups of kernel proteins probably also depends on the bread-making processes used. Therefore, it is necessary to study both genetic and environmental variation in the amount of these other groups of proteins in relation to wheat quality. Because both plant breeders and millers aim at the production of flour suited for as many bread-making processes as possible, the protein composition should preferably be related to several bread-makingprocesses. Another subject requiring additional research is development of small-scale tests, which can be used in breeding programmes to determine the amounts of the HMWg subunits and other proteins affecting wheat quality. These tests could also be valuable for the milling and baking industry, but the requirements of the industry for such a test are not fully congruent with those of the breeders (phenotypic quality in contrast to genotypic quality).The major new results presented are as follows:- The relationship between the type of HMWg subunit and bread-making quality is more complex than generally suggested in the literature.- An efficient method for simultaneous identification and quantification of individual HMWg subunits has been developed.- The amount of the individual HMWg subunits and the proportion relative to the other proteins are subject to environmental and genetic variation. Within genetic variation, effects of HMWg genotype and genetic background can be distinguished.- The ratio between the individual subunits of a genotype is only slightly affected by environmental differences. On this basis, genetic variation in the level of expression of individual genes can be distinguished from environmental variation.- Indications were found that there is a mechanism that coordinates the expression of individual HMWg genes and determines the total amount of the subunits.- The amount of the <em>Glu-B1</em> subunits is positively correlated with the contribution to quality of the alleles encoding these subunits.- The HMWg subunits, i.e. proteins of major importance for the bread-making quality, constitute only a minor fraction of the wheat flour proteins.