The prevalence and impact of Fusarium head blight pathogens and mycotoxins on malting barley quality in UK

Fusarium head blight (FHB) caused by Fusarium and Microdochium species can signi ﬁ cantly affect the yield of barley grain as well as the quality and safety of malt and beer. The present study provides new knowledge on the impacts of the FHB pathogen complex on the malting and brewing quality parameters of naturally infected barley. Quantitative real-time PCR and liquid chromatography double mass spectrometry were used to quantify the predominant FHB pathogens and Fusarium mycotoxins, respectively, in commercially grown UK malting barley samples collected between 2007 and 2011. The predominant Fusarium species identi ﬁ ed across the years were F. poae , F. tricinctum and F. avenaceum . Microdochium majus was the predominant Microdochium species in 2007, 2008, 2010 and 2011 whilst Microdochium nivale predominated in 2009. Deoxynivalenol and zearalenone quanti ﬁ ed in samples collected between 2007 and 2009 were associated with F. graminearum and F. culmorum , whilst HT-2 and T-2, and nivalenol in samples collected between 2010 and 2011 correlated positively with F. langsethiae and F. poae , respectively. Analysis of the regional distribution and yearly variation in samples from 2010 to 2011 showed signi ﬁ cant differences in the composition of the FHB species complex. In most regions (Scotland, the South and North of England) the harvest in 2010 had higher concentrations of Fusarium spp. than in 2011, although no signi ﬁ cant difference was observed in the Midlands between the two years. Microdochium DNA was signi ﬁ cantly higher in 2011 and in the North of England and Scotland compared to the South or Midlands regions. Pathogens of the FHB complex impacted negatively on grain yield and quality parameters. Thousand grain weight of malting barley was affected signi ﬁ cantly by M. nivale and M. majus whilst speci ﬁ c weight correlated negatively with F. avenaceum and F. graminearum . To determine the impact of sub-acuteinfectionsoftheidenti ﬁ ed Fusarium and Microdochium speciesonmaltingandbrewing qualityofnaturally infected samples, selected malting barley cultivars (Optic, Quench and Tipple) were micromalted and subjected to malt and wort analysis of key quality parameters. F. poae and M. nivale decreased germinative energy and increased water sensitivity of barley. The fungal biomass of F. poae and F. langsethiae correlated with increased wort free amino nitrogen and with decreased extract of malt. DNA of M. nivale correlated with increased malt friabilityaswellasdecreasedwort ﬁ ltrationvolume.The ﬁ ndingsofthisstudyindicatethattheimpactofspecies such as the newly emerging F. langsethiae , as well as F. poae and the two non-toxigenic Microdochium species should be considered when evaluating the quality of malting barley.


Introduction
Fusarium head blight (FHB) is an important disease of barley (Hordeum vulgare) caused by a complex of toxigenic Fusarium spp. and non-toxigenic Microdochium spp. known to impact significantly upon the yield and several functional parameters of grain related to malting and brewing quality (Sarlin et al., 2007;Schwarz, 2003; Commission set legislative limits for the main mycotoxins produced by the Fusarium species in cereals and cereal products intended for human consumption (The European Commission 1881/2006. At present, the legislation includes DON and ZON with limits of 1250 ppb and 100 ppb respectively for unprocessed cereals (EC 1881(EC /2006. No legislative limit has been set for NIV as the amount of NIV usually follows closely the levels of DON and thus it is envisaged that the legislation for DON will prevent unacceptable exposure to this toxin (Leslie et al., 2008). New indicative limits for HT-2 and T-2 were published in 2013 as the combined maximum of HT-2 and T-2 toxins of 100 ppb for unprocessed wheat and 200 ppb for unprocessed barley (EC 2013/165/EU).
The immediate effects of severe pre-harvest infection of barley with the species of the FHB complex are reduced seed germination and grain functionality affecting the marketability of the crop and ability to attract malting premium. Further quality problems arise during malting and brewing with severely infected malts being associated with the occurrence of gushing and/or changes in colour and flavour of the finished beer (Oliveira et al., 2012a).
To ensure the quality of barley grain destined for commercial malting and brewing and deemed acceptable for this purpose, the UK malting industry has imposed strict minimum grain specifications which must be met by producers (Assured malt UK, 2008). In addition to quality requirements of acceptable commercial viability of the grain of more than 98% germinative energy (GE), minimum standards include inspection for fungal contamination at intake and due diligence testing for mycotoxins thus preventing heavily infected and quality compromised grain bulks entering the supply chain of malt to beer (HGCA, 2002).
The majority of information on the impact of FHB disease on malting and brewing quality has been provided by artificially inoculated pre-or post-harvest experiments of barley grain and malt using individual Fusarium species. These studies have identified F. graminearum and F. culmorum as the most damaging from the FHB complex, followed by F. poae and F. avenaceum, impacting on several malting and brewing quality parameters (Schwarz et al., 2001). A field experiment using artificial inoculation of barley heads pre-harvest with F. graminearum, F. culmorum or F. poae showed that inoculation resulted in significant reductions in grain plumpness and germination capacity and a slight increase in protein and nitrogen content in the grain (Sarlin et al., 2005). The same three Fusarium species were shown to induce gushing of beer with F. culmorum and F. graminearum being the most potent inducers (Sarlin et al., 2005). Two malting experiments using barley grain artificially inoculated post-harvest with F. culmorum demonstrated an increase in friability, protease and β-glucanase activities, lower amylase activity, greater proportion of free amino and soluble nitrogen and lower β-glucan content, a significant malt loss, as well as a change in the protein content compared to malt from non-infected grain (Oliveira et al., 2012b(Oliveira et al., , 2013. Experiments using artificial inoculation with individual Fusarium species provide valuable information on the severity of FHB related issues in the worst case scenario, but they are less representative of commercial crop production situations where barley grain is likely to be infected by more than one causal organism. Indeed, recent surveys of European commercial barley crops have shown that the FHB complex occurring on the crop is much more diverse than previously considered, including, apart from F. graminearum (F. graminearum sensu stricto), F. culmorum and F. poae, mixed populations of newly emerging pathogens such as F. langsethiae, F. avenaceum, F. tricinctum and Microdochium nivale and Microdochium majus (Nielsen et al., 2011(Nielsen et al., , 2013. Thus, the cumulative impact of the FHB species complex and their related mycotoxins in naturally infected barley upon the malting and brewing quality parameters within limits of acceptable malting capability has not been previously investigated. Furthermore, there is no published information on the effects of F. langsethiae, a potent HT-2 and T-2 producer in barley, or non-toxigenic species such as Microdochium occurring in temperate geographical locations. Here we report on the distribution, co-occurrence and impacts of diverse FHB fungal communities in commercially grown barley crops. Quantitative real-time PCR (QPCR) and LC/MS/MS were applied to quantify pathogen DNA and mycotoxin concentrations, respectively, and a sub-set of the survey samples was subjected to micromalting and laboratory mashing analysis in order to determine potential quality impacts related to naturally occurring mixed fungal loads and mycotoxins.
The present study is based on two annual surveys of commercially grown UK spring malting barley varieties collected in 2010 and 2011, as well as UK spring barley samples collected as part of a previous mycotoxin survey between 2007 and 2009. The main objectives of this study were i) to identify and quantify the main species of the FHB complex and their related mycotoxins in naturally infected field samples of UK malting barley, ii) to determine the regional distribution and co-occurrence of the predominant species associated with FHB disease in UK, iii) to assess the influence of known agronomic factors on fungal populations and iv) to quantify the cumulative impact of fungal and/or mycotoxin contamination on the malting and brewing quality parameters of barley grain as close as possible to commercial malting standards for grain viability.

Malting barley grain samples
A total of 228 samples of malting barley of commonly grown varieties were collected at harvest between 2007 and 2011 from UK fields. Malting barley samples (n = 63) from harvests 2007-2009, which contained a known range of mycotoxins, were obtained from a previous project studying the occurrence of Fusarium mycotoxins in malting barley (Edwards, 2012). During 2010 and 2011 malting barley samples (n = 165) and limited agronomy data including region and barley cultivar, were provided from commercially grown barley fields in UK. A summary of the distribution of the samples included in this study based on sampling year, numbers used in analysis, sampling region and variety is shown in Tables A.1 and A.2.

DNA extraction
Barley grain samples (2 kg) collected at harvest were mixed manually and divided into sub-samples for microbiological, molecular, mycotoxin and brewing analysis. Two hundred grams of each sample was milled (ZM100, Retsch UK Ltd., Leeds, with a 1 mm screen) and stored at − 20°C until DNA extraction. Flour samples (4 g) were weighed individually into 50 ml tubes and 30 ml CTAB buffer (87.7 g NaCl, 23 g sorbitol, 10 g N-lauryl sarcosine, 8 g hexadecyl trimethylammonium bromide, 7.5 g ethylenediamine tetraacetic acid and 10 g polyvinylpolypyrolidone, made up to 1 l with distilled water) was added. The contents were mixed and incubated at 65°C for 2 h. Ten millilitres of 5 M potassium acetate was added to the tubes, which was then mixed and stored at −20°C overnight. Samples were thawed and centrifuged at 3000 ×g for 15 min. A 1.2 ml volume of supernatant was transferred to a sterile 2.0 ml Eppendorf tube, then chloroform (0.6 ml) was added and the contents mixed for 1 min and centrifuged at 11,000 ×g for 15 min. A portion of the aqueous phase (1 ml) was removed and transferred to a new sterile 2.0 ml Eppendorf tube containing isopropanol (0.8 ml), mixed for 1 min and placed for 1 h at − 20°C. The samples were centrifuged at 12,000 ×g for 15 min and the resulting DNA pellets were washed twice with 1 ml of 44% isopropanol. Pellets were air dried and resuspended in 0.2 ml TE buffer and incubated at 65°C for 2 h. The samples were vortexed and centrifuged at 12,000 ×g for 5 min. DNA was measured and quantified based on absorbances at 260 nm, 280 nm, 328 nm and 360 nm using a Cary® 50 spectrophotometer (Varian, CA, USA) and diluted to a working stock of 20 ng/μl and stored at −20°C.

Quantitative real-time PCR
Morphological identification of FHB related species on selected barley samples from 2010 was performed according to the procedures described in the Fusarium Laboratory Manual (Leslie and Summerell, 2006) to determine the most commonly occurring species for later quantification by QPCR.
All malting barley DNA samples were analysed using QPCR to quantify the species of the FHB complex found to be the most frequently occurring in these samples by morphological identification. Amplification and quantification of the relevant species in the malting barley flour samples were performed using a real-time PCR thermal cycler CFX96 (Bio-Rad, UK). Pure DNA isolates, which had been identified and verified by both morphological and molecular methods, of F. graminearum (isolate 212, University of Nottingham), F. culmorum (isolate 236, University of Nottingham), F. avenaceum (isolate 40, University of Nottingham), F. tricinctum (isolate 53, University of Nottingham), F. poae (isolate 246, University of Nottingham), F. langsethiae (isolate 227, University of Nottingham), M. nivale (isolate 226, University of Nottingham) and M. majus (isolate 224, University of Nottingham) were used to make DNA standard curves (10 0 -10 − 6 ng/μl). The amplification mix for each species consisted of 250 nM of each primer (forward and reverse) and 2 × iQ SYBR Green Supermix (Bio-Rad, UK) reagent which was used according to manufacturer's instructions. The volume of DNA sample in the reactions was 2.5 μl in a total reaction volume of 12.5 μl. In the negative control 2.5 μl of PCR-grade water replaced the DNA template. The detection limit for all eight assays is 10 −4 pg/ng total fungal DNA and all assays had an efficiency of E = 95-103%. Species specific primers used for quantification of the species of interest, assay efficiency and references are presented in Table A.3.

Mycotoxin analysis
Mycotoxin analysis on all samples was performed by Campden BRI (Chipping Campden, UK) using UKAS accredited procedures. The trichothecenes (DON, NIV,HT-2 and T-2) and zearalenone were extracted from flour samples (25 g) into an acetonitrile/water mixture with further clean-up of the trichothecenes by solid phase extraction by passing the filtrate through Bond Elut Mycotoxin SPE columns (Agilent Technologies, Germany) (Klotzel et al., 2006). The LC/MS/MS analysis was performed on an Agilent 1200 Infinity LC system (Agilent Technologies, Germany) with a binary pump, coupled with Agilent 6490 MS/MS ESI (Agilent Technologies, Germany) and equipped with an analytical column Agilent Poroshell 120 EC-C18 (2.1 × 100 mm, 2.7 μm, ID Agilent Technologies, Germany). The flow rate was set at 0.2 ml/min and the injection volume was 10 μl. Mobile phase A was water with 0.2% acetic acid and 5 mM ammonium acetate, mobile phase B was methanol with 0.2% acetic acid and 5 mM ammonium acetate. A linear binary gradient was applied from 20 to 70% phase B within 30 min. The content of phase B was then lowered to 20% within a minute followed by equilibration of the column for 10 min. Quantitative determination of all compounds was performed by operating the mass spectrometer in ESI positive and negative ionisation modes (Schuhmacher et al., 2005).
The quantification of the samples was carried out using matrixmatched standards prepared in-house. Spiked samples were included in each batch to determine extraction recovery. The method had an acceptable recovery range for each trichothecene of 60-120%. The results were corrected for recovery.
The expanded measurement of uncertainty was calculated using a standard coverage factor of 2, equivalent to a confidence of approximately 95% in that the actual level of the mycotoxin being measured lies within the quoted range. The expanded measurement of uncertainty was calculated to be 16% for DON and 13% for ZON. The LOQ for the trichothecenes was 10 ppb and for ZON was 2 ppb. Samples below the LOQ were entered as (LOQ) / 2 in the calculation of mean values.

Grain quality assessment
All samples were assessed for the determination of grain quality parameter thousand grain weight (g, TGW) and specific/hectolitre weight (kg/hl, SPW).

Germinative energy (GE)
GE (4 ml) and GE (8 ml) counts were conducted according to European Brewery Convention (EBC) standard methods (Analytica-EBC, Method 3.6.2). Water sensitivity was calculated from the difference between the 4 ml and 8 ml counts, expressed as a percentage. Samples (350 g) were malted in a Custom Lab Micromaltings K steep-germinator and kiln (Custom Laboratory Products, Keith, UK). A manual steeping programme using individual polypropylene tubs was developed so that the steep water was not shared between different samples with different grain microflora. The tubs were floated on the automatically filled steep water in the chamber so that the micromaltings controlled temperature through steeping. Germination and kilning stages were automated. Key process parameters were as follows: steeping: 800 ml of temperate steep water was added to 350 g barley during each steep. Temperature was 16°C throughout and manual water changes were used to create a '3-wet' steep cycle as follows: 8 h wet -16 h dry -8 h wet -16 h dry -2 h wet. Germination: samples were transferred to individual malting 'cages' and germinated at 16°C for 4 days, with automatic turning of the sample cages set at 1 min every 10 min. Kilning: the air on temperature cycle during drying was as follows: 55°C for 8 h, 65°C for 10 h, 75°C for 2 h, and 80°C for 2 h.

Malt analyses
Malt moisture content was measured according to Analytica-EBC, Method 4.2.
Malt friability was measured according to Analytica-EBC Method 4.15 using a Pfeuffer Friabilimeter (Pfeuffer GmbH, Kitzingen, Germany) loaded with 50 g of malt and operated for the standard 8 min. The equipment was calibrated using EBC standard malt samples. Malt α-amylase (dextrinising units, DU) was measured using the Ceralpha Megazyme kit and Malt β-amylase was measured using the Betamyl-5 kit (Megazyme, Bray, Ireland).

Laboratory mashing and associated analyses
Finely ground (0.2 mm screen) malt (50 g) was mashed using a 1-Cube R8 laboratory mash bath (1-Cube, Havlickuv Brod, Czech Republic) according to Analytica-EBC 4.5.1 (congress mash). Laboratory wort filtration volume was measured according to the method of Evans et al. (2011). After returning the first 100 ml of wort collected during laboratory wort filtration, the volume of wort filtered in the next 25 min was measured as an index of mash filterability. The resulting bright worts were analysed for: hot wort extract using an Anton Paar DMA 4500 density metre according to Analytica-EBC Method 4.5.1, free amino nitrogen (FAN) by the spectrophotometric ninhydrin method (Analytica-EBC Method 4.10), wort viscosity according to Analytica-EBC Method 8.4 and EBC wort colour according to Analytica-EBC Method 4.7.1.

Data analysis
All data, apart from malting and brewing data, were analysed using Genstat® Version 14.1 for Windows (VSN International Ltd., UK). Relationships between pathogen DNA and mycotoxins were analysed using single linear regression analysis. Multiple linear regression with groups was used to identify relationships between the DNA of Fusarium spp. and Microdochium spp. and quality parameters of barley grain such as TGW and SW. Where necessary DNA or mycotoxin data were log 10 transformed to normalise residual distributions.
Unbalanced analysis of variance, using linear regression was carried out on fungal and mycotoxin data from 2010 to 2011 to determine the significance (P b 0.05) of sampling region and malting barley variety. It was not possible to include data from 2007 to 2009 in this analysis as samples from these years were not randomly selected but on the basis of their known mycotoxin contents. Therefore descriptive statistics were used for the DNA, mycotoxin and malting/brewing data on all selected samples. The DNA of Fusarium spp. and Microdochium spp. and malting/brewing parameters of samples is presented as mean with 95% confidence intervals and the mycotoxin data is presented as mean with 95th percentile and maximum values. The co-existence of the species of the FHB complex was explored using Principal Component Analysis (PCA) on the correlation matrix of eight variables. These variables were fungal biomass (log 10 pg/ng of total DNA) of F. graminearum, F. culmorum, F. poae, F. tricinctum, F. avenaceum, F. langsethiae, M. majus and M. nivale.
Malting and brewing quality data were entered retrospectively into a D-optimal factorial design space using experimental design software (Design Expert, v 7.0, Stat-Ease, Mn, USA). The malting and brewing quality parameters for the 54 barley samples were entered as responses and modelled against 15 factors, which were: the DNA contents of the individual species analysed in the samples for two Microdochium and six Fusarium species (QPCR data), the barley cultivar, harvest year and the concentrations of five mycotoxins analysed in the samples (NIV, DON, HT-2, T-2, ZON). Modelling progressed via progressive factor reduction, successively removing factors which were of least significance in derived models, until a significant model resulted with factors each of which was significant (P b 0.05) and the model R 2 was maximised. Interactions between factors were included in models where significant.

Fusarium spp. and Microdochium spp. in UK malting barley 2007 to 2011
Species specific QPCR assays were used to quantify Fusarium spp. and Microdochium spp. in UK malting barley samples collected between 2007 and 2011, data presented in Table 1 as mean value with 95% confidence intervals and incidence (%) for each species. When considering the amount of DNA of the eight quantified species of the FHB complex, the non-toxigenic M. majus was the predominant species in samples collected in 2007, 2008, 2010 and 2011 whereas M. nivale was the predominant species in 2009. F. poae was the main Fusarium species in 2007, 2008 and 2009, whereas F. tricinctum predominated in 2010 and F. avenaceum predominated in 2011. The incidence of the species was calculated according to the presence of DNA in all samples throughout the study and the most frequently occurring species in the majority of the analysed samples were F. avenaceum (100%), followed by M. nivale (96%), M. majus (90%) and F. poae (90%). Less frequently occurring species were F. tricinctum (81%), F. langsethiae (65%), F. graminearum (46%) and F. culmorum (36%).

Principal component analysis (PCA)
Quantified DNA of the Fusarium spp. and Microdochium spp. in samples collected in 2010 and 2011 (n = 151) are plotted as a biplot in Fig. 1. This shows both the distribution of the samples in the two most descriptive dimensions of data and the variables (species) projected onto these two axes. On the x-axis, Factor 1 describes 45.91% of the variability and, on the y-axis, Factor 2 describes an additional 15.84% of the original variability. From the principal component analysis, the co-existence of the different species of the FHB complex is visualised in four clusters. The first cluster consisted of M. majus and M. nivale, the second of F. avenaceum and F. graminearum, the third consisted of F. culmorum and F. poae and a fourth cluster consisted of F. langsethiae and F. tricinctum. From the PCA analysis, it is evident that there is a strong association between the occurrences of M. nivale and M. majus and a distinctive negative association between the Microdochium group and the cluster of F. langsethiae and F. tricinctum.

Fusarium mycotoxins in UK malting barley 2007 to 2011
The results from the mycotoxin quantification by LC/MS/MS of a total of 143 samples from 2007 to 2009 and selected samples of 2010 (35) and 2011 (45) are presented in Table 2 as mean value, 95th percentile and maximum value. DON, ZON and NIV predominated in the

Relationships between Fusarium DNA and mycotoxin concentrations
Regressions of mycotoxin concentrations on the quantified Fusarium DNA in the analysed barley samples were carried out to identify the main producers associated with grain contamination. All samples above the limit of quantification of individual mycotoxins by the LC/MS/ MS assay were used in the regression analysis and samples below the limit of mycotoxin quantification were excluded from the analysis. All regressions of mycotoxins on individual or mixtures of species fitted common lines for the data from individual years suggesting that the relationship between Fusarium mycotoxins and their producers is consistent across seasons. A significant positive relationship was observed between the total amounts of F. graminearum and F. culmorum DNA and the amount of DON in the analysed barley grain samples from 2007 to 2009 accounting for 60% of the variance (P b 0.001, R 2 = 0.60, d.f. = 58) ( Fig. 2A). Regressing DNA of individual species accounted for less of the variance, 41% for F. graminearum and 28% of the variance for F. culmorum (data not presented). A similar significant relationship (P b 0.001) was between the total amount of F. graminearum and F. culmorum DNA and ZON accounting for 40% of the variance (data not presented). Analysing F. graminearum and F. culmorum individually showed that both species were equally similarly associated with ZON but accounted individually for only 30% of the variance (data not presented). F. poae DNA showed a significant positive relationship with NIV (P b 0.001, R 2 = 0.84, d.f. = 72) with 73 of the samples from the 2010 to 2011 harvests fitting a common linear regression (Fig. 2B). A significant positive relationship was also found in 2010 between F. langsethiae with total amounts of HT-2 and T-2 (P b 0.001, R 2 = 0.48, d.f. = 15) (Fig. 2C). All positive samples (16) with HT-2 and T-2 above LOQ were included in the regression analysis and all samples contained F. langsethiae.

Effect of cultivar, region and season on the occurrence of FHB pathogens and Fusarium mycotoxins in barley
The regional and seasonal differences in the amounts of total Fusarium DNA and total Microdochium spp. DNA found in UK (South, Midlands, North and Scottish) malting barley samples from 2010 to 2011 are shown in Fig. 3A and B respectively. Significantly higher concentrations of Fusarium species were found in the South of England and in Scotland in 2010, however there were no significant differences between years for the Midlands. In the North of England, Fusarium DNA was found in greater amounts in 2010 than in 2011 (Fig. 3A). The analysis of the regional distribution of Microdochium species showed that Microdochium DNA increased significantly in 2011 in all regions but was found consistently in highest concentrations in the North of England and in Scotland (Fig. 3B).
There were no significant seasonal effects or interactions between season and cultivar. Cv Shuffle was the only variety which contained significantly lower amounts of total fungal DNA compared to cv Concerto, cv Forensic, cv Optic and cv Westminster (P = 0.042, n = 150).

Effect of FHB related species on grain quality parameters
Multiple linear regressions with groups were used to analyse the relationships between grain quality parameters such as thousand grain weight (TGW; g) and specific weight (SW; kg/hl) and the DNA of individual Fusarium and Microdochium species in the collected barley samples from different years. Only grain samples with sufficient grain numbers available for analysis were included in the regression analysis. Regression of TGW (d.f. = 177) on DNA of M. majus, M. nivale and F. avenaceum were significant and fitted separate, non-parallel lines for each season (different slopes and intercepts) accounting for 40% of the variance (Table 3) (different intercepts) for each season (Table 3). The lines were with negative slopes for all seasons, accounting for 48% of the variance.

Malting and brewing quality parameters
A summary of analytical data for the micromalted samples (n = 54) for each barley cultivar, Optic, Tipple and Quench, and season 2010 and season 2011 is presented as mean and 95% confidence interval in Table 4. Cv Optic and cv Quench produced malts with a greater friability than was observed for cv Tipple using the same micromalting programme. Within each cultivar, the friabilities of malts prepared from the 2010 harvest were somewhat higher than in 2011. In accordance with this, malt α-amylase dextrinising units (DU) were higher on average for malts from the 2010 harvest. The laboratory wort filtration volume (ml) followed similar trends in both 2010 and 2011 with the highest volumes obtained when filtering cv Optic worts, followed by cv Tipple and cv Quench. Laboratory wort viscosity (mPa·s) was higher in 2011 than in 2010 for cultivar Tipple only. This is in accordance with the observed lower friability of Tipple malts prepared in 2011. Laboratory wort colour (EBC) followed the same trends in both 2010 and 2011 with cv Quench producing slightly darker wort colours than cv Optic and cv Tipple.

Modelling of malting and brewing quality data
Wort viscosity and malt β-amylase data for the selected 54 samples failed to reveal any significant trends when modelled against the 15 specified factors (season, barley cultivar, DNA of individual species and mycotoxin data). However, for the remaining parameters, significant models were derived which could satisfactorily predict variations in each parameter across the design space (Table 5). The models with the best predictive power (highest model R 2 ; Table 5) were those for water sensitivity of the barley and for laboratory wort colour. The DNA of the individual species identified as significant model terms (Table 5) were those for F. poae (GE (4 ml), water sensitivity, laboratory wort extract and wort FAN), F. langsethiae (laboratory wort extract and wort FAN) and M. nivale (GE (4 ml), water sensitivity, malt friability, laboratory wort filtration volume and laboratory wort colour). The directionality of the factor effects is indicated in Table 5 with a (+) or (-). For example, the increased presence of pathogen DNA for F. poae and M. nivale correlated with a reduction (-) in GE (4 ml) counts and an increase (+) in water sensitivity. Model data for laboratory wort colour is shown in Fig. 6. Laboratory wort colour was the best fitting predictive model with a reasonable correlation between predicted and actual wort colour values. The factor plots indicate that the action of increasing M. nivale concentration was to increase laboratory wort colour by approximately 1 EBC colour unit across the range of concentrations encountered (Fig. 6B) whilst there were significant differences between the barley cultivars Quench and Tipple for wort colour (Fig. 6C).
The analytical concentrations of mycotoxins (NIV, DON, ZON, HT-2 and T-2) and the species DNA data of certain species (F. avenaceum, F. tricinctum, F. graminearum, F. culmorum, M. majus) were not found to be significant factors in any of the models developed. Of the mycotoxins, NIV was the closest to approaching significance (P b 0.05) in some models. However it co-varied closely with F. poae (the main NIV producer) and the models were, in each case, stronger when modelled against F. poae DNA rather than the NIV concentrations.

The occurrence of FHB pathogens in UK malting barley
This is the first study using commercially grown, naturally infected malting barley to investigate the cumulative impact of diverse populations of FHB pathogens and their mycotoxins on malting and brewing quality parameters. The findings show that the naturally occurring composition of species of the FHB complex on malting barley in UK is diverse and dominated by non-toxigenic Microdochium species and the toxigenic Fusarium species F. poae and F. avenaceum. The presence and amount of species DNA showed yearly variation. M. majus was the dominant species in 2007, 2008, 2010 and 2011 and M. nivale in 2009. Relatively lower amounts of F. langsethiae, F. graminearum and F. culmorum were found in all five years. Similarly, a survey of the composition of the FHB complex in Danish barley, grown under similar environmental conditions as in the UK, between 2005 and 2007 found that M. majus was the predominant species followed by F. langsethiae, F. avenaceum and F. poae (Nielsen et al., 2011(Nielsen et al., , 2013. The predominance of F. avenaceum in barley has also previously been shown in Finnish barley . F. graminearum is typically considered to be the most prevalent and aggressive FHB pathogen on both wheat and barley in much of the world, particularly in the temperate and warmer regions of the USA, China and the southern hemisphere, whereas F. culmorum was associated   (Osborne and Stein, 2007). However, our findings together with the recent work in Europe (Nielsen et al., 2011(Nielsen et al., , 2013Yli-Mattila et al., 2008) strongly suggest that F. graminearum and F. culmorum are not the most important pathogens, and particularly F. culmorum is occurring less, as part of the FHB complex in barley. This is of particular importance in European locations where research focus should be directed towards understanding the impact of other species previously considered less aggressive but still economically important due to their association with FHB disease and mycotoxin accumulation. PCA identified clear groupings of co-occurring pathogen species (Fig. 1). Similar to previous studies (Nielsen et al., 2013), F. culmorum was found to associate more closely with F. poae, whereas a negative association was found between the cluster of F. langsethiae and F. tricinctum and the cluster of Microdochium species. F. avenaceum co-existed together with F. graminearum and multiple regression analysis showed that both species negatively influenced pre-harvest quality factors of the crop such as specific weight. Furthermore, the non-toxigenic Microdochium species, which were found to strongly co-exist were also found to impact on the yield parameter TGW. To our knowledge, there are no previous reports on the effects of mixed populations of Fusarium and Microdochium spp. on yield parameters of malting barley.

Agronomy impact on FHB complex composition
Significant differences between regions and years for species composition were evident (Figs. 3 & 4), with higher concentrations of Fusarium spp. in the South and North of England and in Scotland in 2010 whilst no significant difference was observed in the Midlands between the two harvests. Analysis of the regional distribution of the two Microdochium species showed that the amount of Microdochium DNA was significantly higher in 2011 than in 2010 and significantly more M. nivale and M. majus were found in the North of England and in Scotland compared to the South or Midlands regions. The regional differences in the species composition of the FHB complex are possibly explained by the differences in their environmental requirements for growth and infection and climatic differences in the different regions. For example, in the North of England and Scotland average temperatures during the active growing season of the spring crop remain below 15°C which is a more favourable environment for growth and infection by Microdochium species (Parry et al., 1995;Xu et al., 2008). In contrast, F. poae requires dry and warm conditions of around 25°C for optimum growth (Parry et al., 1995;Xu et al., 2008). F. graminearum infection is more often associated with wet and warm conditions during anthesis, whereas F. culmorum, F. avenaceum and F. tricinctum require wet, humid and cool environmental conditions (Xu et al., 2008).
There were only small differences between the barley cultivars included in our studies with respect to the amounts of pathogen DNA present. The exception was cv Shuffle which had significantly lower amounts of total fungal DNA, irrespective of region, compared with the other elite cultivars such as Concerto, Forensic, Optic, Westminster (P = 0.042). This indicates that current commercially available cultivars, at least in the UK, are of similar susceptibility to Fusarium infection. Only a few sources of FHB resistance are known in barley, however, the level of resistance, even in these, is at best moderate (Bai and Shaner, 2004).

Mycotoxin contamination and safety of English and Scottish malting barley
Mycotoxin analysis of the UK barley samples revealed that the predominant mycotoxins were DON followed by NIV and ZON and lastly by HT-2 and T-2 at low concentrations. In 2010 and 2011 a large number of samples were analysed to obtain a representative overview of the natural mycotoxin contamination in English and Scottish fields and these were all found to be below the legislative limits of Fusarium related mycotoxins. In contrast to HT-2 and T-2, DON and NIV were found in significantly higher concentrations in 2011 than in 2010. The sum of HT-2 and T-2 found in the barley samples from 2010 was significantly associated with DNA of F. langsethiae. Besides F. langsethiae, F. sporotrichioides is also known to produce HT-2 and T-2 (Thrane et al., 2004). However in the UK, previous studies in oats have shown a strong relationship between combined HT-2 and T-2 levels and DNA amounts of F. langsethiae , whereas in Europe three different species, F. langsethiae, F. sporotrichoides or Fusarium sibiricum, are associated with HT-2 and T-2 (Fredlund et al., 2010;Yli-Mattila et al., 2008). The barley samples were analysed for F. sporotrichioides DNA with primers known to cross-react with F. sibiricum (Yli-Mattila et al., 2011) but failed to detect the DNA of either species or to isolate any of these species from barley grain. Thus, the evidence suggests that in the UK, contamination with HT-2 and T-2 in both oats and barley is predominantly associated with F. langsethiae.
Isolates of F. graminearum, F. culmorum and F. poae are able to produce NIV; in the present study only F. poae correlated strongly (R 2 = 0.84) with NIV concentrations. This finding is in agreement with previous studies in Northern Europe showing that F. poae is the main producer of NIV in barley (Bushnell et al., 2003;Nielsen et al., 2011;Yli-Mattila et al., 2004, and similar trends have been observed in UK oats . Samples collected from 2007 to 2009 were limited and not fully representative of these harvest years. These samples were selected on the range of their known mycotoxin concentrations and used to isolate, identify and quantify the main producers associated with mycotoxin accumulation. Of these samples only a single sample exceeded the legislative limit set for DON, a total of 19 samples of the 63 exceeded the legislative limit of ZON and none of the analysed samples exceeded the indicative limit of HT-2 and T-2. Regression analysis using individual DNA of F. graminearum or F. culmorum revealed that the relationship with DON and ZON was stronger for F. graminearum. However the regression for producers and mycotoxins was fitted best with cumulative data including F. culmorum DNA, suggesting that both species were implicated in the production of DON and ZON. The predominance Table 4 Summary data for malt and wort quality parameters, based on a total of 54 micromalted barley samples and broken down according to barley variety and season described by mean and 95% confidence interval. Abbreviations: number of samples (n) and germinative energy (GE). of F. graminearum and F. culmorum as the main DON producers agrees with previous reports of strong correlations between the DNA of DONproducing Fusarium species and DON concentrations found in barley, wheat and oats (Fredlund et al., 2008;Nielsen et al., 2011;Sarlin et al., 2006;Waalwijk et al., 2004;Yli-Mattila et al., 2008. In two samples, low levels of DON were detected but the DNA of F. graminearum or F. culmorum were below the level of quantification, thus it is possible that DON in these two samples may have been associated with different producers, for example Fusarium equiseti and/or Fusarium acuminatum which are also known DON producers in cereals in Europe (Marín et al., 2012).

Impact of FHB pathogens on malting and brewing quality parameters
The major species to significantly affect quality parameters in further micromalting studies were identified as M. nivale, F. poae and F. langsethiae. Although the survey samples from 2010 to 2011 were randomly selected from sites across the UK (and thus representative), it should be noted that barley with GE (4 ml) counts of less than 98% would not be processed in commercial malting. In the present experiment, the GE (4 ml) requirement was relaxed to N80%, in order to include samples in the survey which had more widely ranging contents of the Fusarium and Microdochium species investigated. These samples would not normally have been malted for brewing use. However, in the majority of cases, the resultant malt friability and α-amylase values confirmed that samples had malted satisfactorily. Thus, the GE of these samples had declined through storage, but the germinative capacity (percentage of live grains) was still sufficient.
The micromalting experiment was not balanced in terms of the numbers of each cultivar malted in each year. This was largely due to the availability of samples within the survey which had sufficient germinative energy to malt and which showed interesting variations with regard to their measured concentrations of fungal DNA and mycotoxins. In general the malts prepared were of acceptable specification (although precise requirements depend on the end user). If anything, the majority of malts were rather well modified (friability N90% and with high α-amylase activities), which was a result of the generous 50 h steep cycle, designed to ensure that barley samples of differing provenance would all hydrate and modify sufficiently.
Water sensitivity is defined as the difference between the GE (4 ml) and GE (8 ml) counts. The number (expressed as a percentage) indicates whether a malt sample has lower germinative energy in the presence of excess water. In the present study, both M. nivale and  F. poae were significant factors which correlated positively with water sensitivity. Crop year was also a significant factor in determining water sensitivity, with 2011 samples having on average, greater water sensitivity than those from 2010. Water sensitivity is of commercial significance because the maltster will need to adjust the steeping process (e.g. the duration of air rests) when malting water sensitive grain. Water sensitivity has been linked to malt microflora (Woonton et al., 2005) although other factors seem to be involved, as treatment of grains with anti-microbial agents does not consistently overcome water sensitivity (Kelly and Briggs, 1992). The fact that water sensitivity was also affected by crop year could be caused by differences in climatic/ agronomic influences during the respective years. It could also reflect the fact that on average more fungal DNA was found in the 2011 samples for the two species identified as being significant in the model for water sensitivity (0.027 pg/ng as compared with 0.015 pg/ng for F. poae and 0.37 pg/ng versus 0.19 pg/ng for M. nivale).
There was a positive correlation of F. poae with wort FAN suggesting that F. poae contributes to proteolytic activity through the malting and mashing processes, thus increasing FAN production, particularly during the low temperature stand at 45°C during the congress mash schedule. The model for wort FAN also included F. langsethiae and an interaction term between the two species. The interaction indicated that at low concentrations of F. langsethiae, F. poae dominated with regard to increasing wort FAN, whereas at high F. langsethiae concentrations and low F. poae, the contribution to FAN from F. langsethiae was significant. The trends found in the interactions of F. poae, F. langsethiae and wort FAN may reflect competitive aspects between the growth habits of these two species. These results are consistent with prior reports of protease secretion by F. poae (Pekkarinen et al., 2000;Schwarz et al., 2002). Pekkarinen et al. (2000) reported that F. poae grown on barley grains produced non-specific protease activities at pH 5.0 (close to mash pH) and pH 8.0, over a period of days following inoculation. These activities were not present in non-inoculated barley. Schwarz et al. (2002) conducted a glasshouse trial where barley plots were inoculated separately with F. graminearum and F. poae. The high wort FAN contents reported for the inoculated plots led the authors to conclude that Fusarium spp. contributed exoproteinase as well as endoproteinase activities.
The results presented here suggest that M. nivale can have a significant impact upon the quality of malting barley. On balance, these impacts were undesirable as, although positively correlated with friability, M. nivale also correlated with increased water sensitivity, lower germinative energy and had a negative impact on the laboratory wort filtration volume. The latter is a crude predictor of the mash separation performance of malt in a brewhouse (Evans et al., 2011). A lower volume of filtered wort after the specified time interval indicates that the mash might take longer to filter on a commercial scale. Although the model for wort filtration volume was significant, it had a low predictive power, indicating that many other variables not accounted for in the present study can influence mash separation performance. M. nivale occurrence, or prevalence in the FHB complex, has been associated with regions experiencing relatively cool temperatures and frequent, short, showers (Doohan et al., 2003;Nielsen et al., 2011).
The absence of any direct relationship between the presence of Fusarium spp. and Microdochium spp. and wort viscosity was contrary to previous reports of a reduction in wort viscosity in Fusariuminfected malts, which was attributed to glucanase and xylanase activities of Fusarium spp. (Schwarz et al., 2002). However, a recent study reported increases in wort β-glucans when brewing with malts prepared from grain artificially inoculated with F. culmorum (Oliveira et al., 2012a). Hence the precise impact of infection may depend on the particular β-glucanase activities present and the mashing schedule employed. β-glucan solubilase activity will solubilise high molecular weight β-glucans during mashing, thus tending to increase wort viscosity. Endo-β-glucanase activities then reduce the mean molecular weight of glucans present and thus act to reduce wort viscosity. It is further true that in most prior studies control malts were compared with artificially inoculated barley malts, whereas in the present trial we investigated natural variations in the grain microflora from survey sites across the UK.
Wort colour is determined to a large extent during kilning. Since the same kiln temperature cycle was used for all samples, colour differences were caused principally by variations in concentrations of the Maillard browning reaction precursors (reducing sugars and free amino nitrogen) present following germination. It is consistent in this regard that increasing M. nivale DNA correlated both with increasing malt friability and laboratory wort colour, since the release of amino acids and reducing sugars from the breakdown of protein and starch respectively, increases with the extent of modification and friability of malt.
There have been several reports on the changes in the diversity of composition of the FHB complex in cereals within different geographical and climatic locations. In the past ten years in Europe, F. culmorum has been replaced by F. graminearum (Waalwijk et al., 2003;Jennings et al., 2004;Stepien et al., 2008;Xu et al., 2008;Isebaert et al., 2009;Nielsen et al., 2011) and furthermore F. poae has been shown to replace F. graminearum in Southern Europe (Pancaldi et al., 2010;Shah et al., 2005). In contrast, in Central Europe and in North America and China, DON is the main trichothecene associated predominantly with F. graminearum and species of the F. graminearum clade (O'Donnell et al., 2004).
In these studies we describe the impact of newly emerging species of importance, M. nivale and F. langsethiae, on the malting and brewing quality of naturally infected barley. The results clearly indicate that the pathogen populations of the FHB complex in barley in the UK are diverse and dominated by non-toxigenic Microdochium species and toxigenic Fusarium species such as F. poae, F. avenaceum and F. langsethiae. Future research efforts should focus on elucidating the impact of these newly emerging species and their mycotoxins, for example, enniatins produced by F. avenaceum and F. tricinctum on barley and HT-2 and T-2 produced by F. langsethiae. Appendix A1 Appendix A2 and T-2 + HT-2 in UK malting barley samples collected in 2010 and 2011. Significant regional differences in NIV (P b 0.001) and DON (P = 0.024). No significant regional difference in T-2 + HT-2 (P = 0.671).