Rothamsted Repository Download

Dietary fibre (DF) has multiple health benefits, and wheat products are major sources of DF for human health. However, DF is depleted in white flour, which is most widely consumed, compared to wholegrain. The major type of DF in white wheat flour is the cell wall polysaccharide arabinoxylan (AX). Previous studies have identified the Chinese wheat cultivar Yumai 34 as having unusually high contents of AX in both water-soluble and insoluble forms. We have therefore used populations generated from crosses between Yumai 34 and four other wheat cultivars, three with average contents of AX (Ukrainka, Altigo and Claire) and one also having unusually high AX (Valoris), in order to map QTLs for soluble AX (determined as relative viscosity) of aqueous extracts of wholemeal flours) and total AX (determined by enzyme fingerprinting of white flour). A number of QTL were mapped, but most were only detected in one or two crosses. However, all four crosses showed strong QTLs for high RV/total AX on chromosome 1B, with Yumai 34 being the increasing parent, and a KASP marker for the high AX Yumai 34 allele was validated by analysis of high AX lines derived from Yumai 34 but selected by biochemical analysis. A QTL for RV was mapped on chromosome 6B in Yumai 34 x Valoris, with Valoris being the increasing allele, which was consistent with the observation of transgressive segregation for this trait. The data indicate that breeding can be used to develop wheat with high AX fibre in white flour.


Introduction
Dietary fibre (DF) is essential for human health, with cereals providing about 40% of the total fibre intake in Western European countries such as the UK (Bates et al., 2014) and Finland (Helldan et al., 2012). Dietary fibre (DF), and wholegrain cereal fibre in particular, has been shown to have a number of health benefits, including lowering blood pressure and serum cholesterol, improving insulin sensitivity and reducing the incidence of certain types of cancer, notably bowel and breast cancers (Cade et al, 2007;Wood, 2007;Anderson et al, 2009;Wolever et al, 2010;Aune et al, 2011;2016;Cooper et al., 2015;Ye et al., 2012;Hajishafiee et al., 2016;Reynolds et al., 2018) The mechanisms are still incompletely understood,but are considered to include increasing faecal bulk and reducing intestinal transit time, binding cholesterol and carcinogens, reducing the rate of digestion and glucose release in the small intestine and fermentation to beneficial short chain fatty acids in the colon. DF also occurs in soluble and insoluble forms, which are considered to differ in some respects in their health benefits, with insoluble fibre being more slowly fermented and contributing particularly to binding cholesterol and carcinogens and increasing faecal bulk.
However, despite these established health benefits, the intake of DF in most countries falls far below the recommended levels. For example, the daily intake in the UK is currently 17.2g/day for women and 20.1 g/day for men, compared with a recommended intake of 30g/day (https://www.nutrition.org.uk/nutritionscience/nutrients-food-and-ingredients/dietaryfibre.html).
Although wholegrain wheat is relatively rich in fibre, containing about 11 to 15% dry weight, (Andersson et al., 2012), most wheat products are made from white flour (derived from the starchy endosperm) (Steer et al., 2008) which contains only 2-3% fibre (Gebruers et al, 2008). Furthermore, increased consumption of highly refined cereal products (including bread and other products from white flour) is occurring in countries undergoing urbanisation and industrialisation, which is considered to contribute to increases in obesity and chronic diseases in these countries (Mattei et al, 2015).
However, the content of AX also varies between different genotypes of wheat. For example, 2-fold variation in the content of total (TOT)-AX and 4.7-fold variation in water-extractable (WE)-AX was reported in white flour of 150 wheat genotypes grown together on a single site (Gebruers et al., 2008), and 2.9-fold variation in WE-AX and 1.7-fold variation in water-unextractable (WU)-AX in 20 wheat cultivars (Ortiz-Ordaz and Saulnier, 2005). Furthermore, a high proportion of the variation in the AX content of wholemeal and white flours of wheat is heritable, and hence accessible for exploitation by breeders (Martinant et al., 1999;Shewry et al., 2010).
However, the exploitation of this variation to develop improved wheats has been limited by the lack of tools for selection, with biochemical analyses being slow and costly and a lack of molecular markers for selection.
A number of studies of the genetic control of AX content have been reported using genetic analysis, with most analysing wholemeal flour by either direct determination of AX (by monosaccharide analysis or colorimetric determination) or the relative viscosity of aqueous extracts as a proxy (Martinant et al. 1998;Perretant et al. 2000;Laperche et al. 2007;Quraishi et al. 2009;Charmet et al. 2009;Nyugen et al;Yang et al., 2015), for AX content. In addition, two association studies of AX in wholemeal tetraploid wheat (Marcotuli et al., 2015) and in white flour of bread wheat (Quraishi et al., 2011)  Although these analyses were carried out on grain samples from single plots grown on the same site in 2005-6, further comparative analyses carried out on lines grown for over 10 years has confirmed that Yumai 34 always contains the highest, or . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/705343 doi: bioRxiv preprint first posted online Jul. 17, 2019; among the highest, contents of both TOT and WE-AX fractions in flour (authors' unpublished results). This is illustrated by Table 1, which compares the contents of AX fractions in white flour of four cultivars grown at Rothamsted in 2011/2012. Yumai 34 clearly has the highest contents of TOT-AX and WE-AX, with the latter being reflected in the high relative viscosity (RV) of aqueous extracts (as discussed below).
Arabinoxylan comprises a backbone of β -D-xylopyranosyl residues linked through (1→4) glycosidic linkages with some residues being substituted with α -Larabinofuranosyl residues at either position 3 or positions 2 and 3. Monosaccharide analysis showed that the ratio of arabinose to xylose (A:X) in TOT-AX is lower for Yumai 34 than for other cultivars, indicating that the structure of AX also differs (Table 1). This difference was therefore investigated further using enzyme fingerprinting. Treatment of AX with a specific type of endoxylanase enzyme results in cleavage of the xylan backbone, releasing a mixture of xylose, xylobiose, xylotriose (comprising 1, 2 and 3 xylose units, respectively) and arabinoxylan-derived oligosaccharides (AXOS) comprising 4 to 7 xylose units, one or more of which may be mono-or disubstituted with arabinose. The proportions of these AXOS therefore provides a "fingerprint" which reflects differences in the extent and pattern of arabinose substitution. The application of this approach to AX from white flours of the four cultivars showed that Yumai 34 has a high proportion of a pentasaccharide with the structure XA3XX (xylose-xylose monosubstituted with arabinose-xylose-xylose), which comprised 26% of the total fragments compared with 15.17%, 16.14% and 17.21% in the other cultivars (Supplementary Table S1). This increase is associated with reduced proportions of oligosaccharides containing disubstituted arabinose residues, and accounts for the lower A:X ratio shown in Table 1.

Development of populations for genetic analysis of AX
Four populations were generated from crosses between Yumai 34 and European cultivars: with the central European cultivar Ukrainka (96 F6 recombinant inbred lines (RILs)) (Y34Ukr), the UK biscuit-making cultivar Claire (95 RILs) (Y34Cl) and the French breadmaking cultivars Altigo (245 doubled haploid lines (DHL)) (Y34Alt) and Valoris (84 DHL) (Y34Val). Whereas Ukrainka, Altigo and Claire and were selected as parents because they had average contents of AX, previous studies have shown that Valoris has higher than average contents of both WE-AX and TOT-AX (0.8% . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/705343 doi: bioRxiv preprint first posted online Jul. 17, 2019; and 2.2% compared with means of 0.51% and 1.93%, respectively, in the Healthgrain study) and it has therefore been used as a "high AX parent" in crosses with other low AX genotypes (Charmet et al., 2009).

Mapping relative viscosity (RV) of aqueous extracts of wholemeal
Most previous genetic analyses of AX fibre in wheat have determined the RV of aqueous extracts of white or wholemeal flours as a proxy for arabinoxylan content (reviewed by Quraishi et al., 2010;Shewry, 2013). This is because this parameter is more readily determined than AX, eliminating the need for milling and biochemical analysis. It is considered to be valid because WE-AX is known to be the major contributor to the viscosity of aqueous extracts. However, the use of RV of wholemeal flour as a proxy for WE-AX in white flour has not been validated. We therefore compared the RV of aqueous extracts and the contents of WE-AX in wholemeal and white flours from 10 lines from the Y34Ukr cross (grown at Where the QTL 1 LOD confidence intervals substantially overlap (green line on figure) for the same trait in different populations we have made the assumption that the underlying effect is the same. The increasing alleles for these QTL come from Yumai 34 except for a QTL for RV on 1A in Y34Val and a QTL for RV on 6B in Y34Val.
A major objective of this study was to understand the genetic basis of the high RV of Yumai 34 so QTL with increasing alleles from this variety, of large effect, and . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/705343 doi: bioRxiv preprint first posted online Jul. 17, 2019; expressed in multiple populations are of particular interest. In this respect 1BL stands out at LOD 12.6 and 7.8 in the Y34Alt and Y34Val populations, respectively, accounting for 16.3% and 24.2% phenotypic variance. With additive effects of 0.53 and 0.99 RV units these QTL alone can deliver substitution effects of 1 to 2 RV units in populations with mean RVs of 3 and 4.5. A weaker RV effect also detected in two populations, Y34Alt and Y34Cl, is found on 3B. No other RV effects are detected in more than one population. Mapping TOT-AX determined by fingerprinting Analyses of ten Y34Ukr lines (discussed above) showed that RV of aqueous extracts  Table 2 or Figure 2 because the LOD scores were below the cut-off. These were a QTL for TOT-AX in Y34Val (LOD 2.5) and a QTL for RV in Y34Ukr (LOD2.3). Interestingly, a second TOT-AX QTL from Y34Ukr is located on the opposite end of chromosome 1B (1BS).
. CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/705343 doi: bioRxiv preprint first posted online Jul. 17, 2019; Three QTL were identified in which Yumai 34 was not the increasing parent, for TOT-AX and RV from Y34Ukr and Y34Val, respectively, which co-located on chromosome 1A and for RV on chromosome 6B in Y34Val only.
Identification and validation of markers for the high AX QTL from Yumai 34 The QTLs identified on 1BL for which Yumai 34 alleles increased TOT-AX and/or RV

Discussion
We have used four crosses with the high fibre wheat cultivar Yumai 34 to identify QTL for high RV and total AX fibre in white flour. Although a number of QTL were mapped, which is consistent with earlier studies, most of these were only detected in one or two crosses. However, all four crosses showed strong QTLs for high AX/RV on chromosome 1B, with Yumai 34 being the increasing parent, although this mapped to the 1BS in the Y34Ukr and to 1BL in the other crosses. Furthermore, a . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/705343 doi: bioRxiv preprint first posted online Jul. 17, 2019; KASP marker for the high AX Yumai 34 allele was validated by analysis of high AX lines derived from Yumai 34 but selected by conventional biochemical analysis. Previous studies have also shown QTLs for RV on chromosome 1B, including a QTL on 1BL with a heritability of 27.4% in one population (Quraishi et al., 2011). This may correspond to the 1BL QTL in Yumai 34, with the Yumai allele giving higher levels of AX and RV. The two QTL on 1B described here appear to be genuinely independent.
We investigated whether a translocation present in Ukrainka might cause 1BL markers in Y34Ukr to link with markers on 1BS, but there was no evidence for this (data not shown). Moreover, the BA00789946 KASP assay is assigned to 1BL when mapped using the Y34Ukr population, as in the other three crosses. Nevertheless, this marker has strong predictive power for high fibre in the progeny selected from crosses between Yumai 34 and Ukrainka, which supports the QTL for RV on 1BL in this population even though the LOD score was below the cut-off.
The most likely explanation for the identification of a TOT-AX QTL on 1BS in the cross with Ukrainka is that Yumai 34 does indeed carry AX increasing effects at both ends of chromosome 1B. However, detecting linked QTLs requires higher statistical power and it is likely that we did not detect the TOT-AX QTL on 1BL as significant in the same cross due to insufficient power.
It is notable that only one cross, between Yumai 34 and the high AX cultivar Valoris, showed significant transgressive segregation. Previous analysis of a cross between Valoris and Isengrain (which has a normal AX level) identified a QTL on 6BL which explained 58.4% of the heritability (Charmet et al., 2009;Quraishi et al., 2011). This may correspond to the QTL identified on 6B in the Yumai 34 x Valoris cross, although this QTL was mapped to 6BS not 6BL. The presence of the high AX allele of the 1B QTL in Valoris probably accounted for the presence of transgressive segregation, which indicates that it should be possible to stack the 1B and 6B QTLs, and possibly also other high AX QTLs.

Experimental procedures
Production and growth of materials  Table 1 and representative enzymatic fingerprinting data in Supplementary Table S1.
High fibre lines selected from a cross between Yumai 34 and Ukrainka were grown at the Centre for Agricultural Research, Martonvásár.

Milling
White flour was produced using a Chopin CD1 mill. Grains were brought to room temperature and moisture content determined using Bruker Minispec mq-20 NMR analyser using an in-house developed calibration. 50g of grain was conditioned to 16.5% moisture overnight prior milling. First break and first reduction flours were combined to give the white flour fraction.

Supporting information
Additional Supporting Information may be found online in the supporting information tab for this article: Table S1. Relative amounts of AXOS and GOS released by enzyme digestion of white flour from four wheat cultivars . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/705343 doi: bioRxiv preprint first posted online Jul. 17, 2019; Figure S1. Correlation Matrix of 10 lines from the Yumai x Ukrainka population (2013/2014) grown at Rothamsted Research of relative viscosity, water extractable (WE) and total AX measured colorometrically (pentosan assay), and total AX by enzyme fingerprinting in wholemeal and white flour.  was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/705343 doi: bioRxiv preprint first posted online Jul. 17, 2019;   Each hash mark on the left-hand scale represents 1Mb. Approximate location of centromeres is shown as a hour glass when determined using chromosome arm survey sequence data (International Wheat Genome Sequencing Consortium, 2014) and as black blocks for 2B and 3B using density of annotated genes (Appels et al 2018). QTL are named as Population-environment-trait (see Table 2). . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/705343 doi: bioRxiv preprint first posted online Jul. 17, 2019; . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/705343 doi: bioRxiv preprint first posted online Jul. 17, 2019; and as black blocks for 2B and 3B using density of annotated genes (Appels et al 2018). QTL are named as Population-environment-trait (see Table 2).
. CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which . http://dx.doi.org/10.1101/705343 doi: bioRxiv preprint first posted online Jul. 17, 2019;  . CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.