Modelling the colonisation of sorghum grain by the Fusarium graminearum species complex and concomitant mycotoxin production

English: Sorghum is the fifth most grown cereal worldwide, and is a staple food in 30 countries that sustains 500 million people in the semi-arid tropics. Sorghum grain mold (SGM) is the one of the most important pre-harvest biotic constraints in sorghum production. Over 40 genera of pathogenic fungi occur on sorghum grain and cause SGM. Fusarium graminearum, a causal agent of SGM, is responsible for the majority of economically and medically important mycotoxins associated with the disease. This, however, is not an individual pathogen, but a complex of species or a combination of related Fusarium species and is referred to as the Fusarium graminearum species complex (FgSC). The FgSC formed the focus of this study. Grain from nine sorghum cultivars, harvested over three seasons at Cedara and two seasons at Alma, was evaluated for grain mold severity, mycotoxin contamination and the stability of the grain mold response over changing environmental conditions. Lower visual threshed grain disease ratings and total fungal biomasses, determined using ergosterol analysis, were observed in grains with an elevated tannin content. However, no correlation between threshed grain disease ratings and ergosterol content was observed indicating that the former criterion is not a reliable measure of grain colonisation by grain mold fungi. Quantitative PCR analysis indicated a FgSC DNA content in grain over a range of 5.52 ng.μl-1 in PAN8625 to 55.43 ng.μl-1 in PAN8806 with significant differences between cultivars. Only three of the 162 grain samples had deoxynivalenol (DON) concentrations that exceeded 10 μg.kg-¹ and DON was therefore excluded from further analysis. However, nivalenol (NIV) and zearalenone (ZEA) were present in all but four and two samples, respectively. Additive Main Effect and Multiplicative Interaction (AMMI) analysis of FgSC DNA, NIV and ZEA concentrations indicated a relatively stable response in cultivars to changing environments with most cultivars yielding an IPCA1 score <1. Robust regression was applied to quantify the relationship between NIV and ZEA accumulation in grain relative to the FgSC DNA concentration and indicated that host genotype influences mycotoxin production despite similar colonisation levels. Results indicate the need for the inclusion of environmental variation in the screening and selection for resistance to SGM in sorghum genotypes, to ensure quality grain and human and animal health. The development of an epidemiological model which quantifies the risk of grain molds and mycotoxins in sorghum production areas could enable producers to ensure that timely management decisions are made to reduce FgSC infection and mycotoxin contamination. Sorghum grain collected over two seasons from 18 South African sorghum production areas were analysed for FgSC colonisation and DON, NIV and ZEA contamination. FgSC colonisation and concomitant mycotoxin accumulation coincided with weather conditions during early-post flowering, 82-95 days after planting (d.a.p.) and soft dough stage, 92-115 d.a.p., which are the critical periods for grain colonisation and mycotoxin accumulation. FgSC development and colonisation were significantly, positively correlated with maximum relative humidity 82-95 d.a.p. and significantly inversely correlated with maximum temperature and evapotranspiration 82-95 d.a.p. DON, NIV and ZEA accumulation were significantly positively related to FgSC DNA concentration. DON had borderline significant positive relationship with maximum temperature 101-115 d.a.p., however NIV and ZEA had significant inverse relationship with minimum temperature 91-104 and 100- 113 d.a.p., respectively. Preliminary models based on stepdown multiple regression analysis were developed. Future studies could include localities with more available and accurate weather data to further calibrate and validate the models developed. A range of commercial sorghum grain samples were collected from a sorghum processing company as well as two finished products were taken from storage. In addition a total of 180 sorghum grain samples consisting of four cultivars from three localities in South Africa (Greytown, Standerton and Potchefstroom) were collected during one production season and decorticated using a tangential abrasive dehulling device (TADD) for five time periods. Ergosterol concentrations were highest in sorghum bran and lowest in 22% dehulled grain, indicating that a high proportion of fungal contamination lies in the outer grain layers. In contrast, FgSC DNA concentrations were detected at lowest levels in sorghum bran and at highest levels in 22% dehulled grain. The assumption was therefore, made that FgSC infections were deep-seated within the grain endosperm. Furthermore, controlled decortication by TADD only resulted in a significant reduction in FgSC DNA content after four minutes i.e. 35% decortication. The assumption is further supported by NIV concentrations in both commercial and controlled samples which were relatively high in sorghum bran and 4% dehulled grain, but reduced to 0.00 μg.kg-1 in both the 22% dehulled grain and grain from the six minutes decortication. In contrast with NIV, ZEA and DON, were removed from grain by short periods of decortication. The assumption can be made that ZEA and DON are associated with superficial FgSC infections and accumulate in the outer layers of the grain while NIV is associated with pathogenesis in the deeper endosperm layers. Cultivars differed in hardness and variation in hardness of grains was associated with prevailing weather conditions at each locality, as well as endosperm texture with lower colonisation and mycotoxin levels in the harder grain. The highest FgSC DNA concentrations and DON, NIV and ZEA accumulation was recorded in grain from Greytown where weather conditions during the critical grain development stages promoted infection and the contamination of grains by mycotoxins. Understanding the effects of decortication on FgSC DNA concentrations and the accumulation of DON, NIV and ZEA could assist commercial processors to make the best management decisions for the removal of these harmful mycotoxins.