Journal of Applied Botany and Food Quality 88, 228 - 233 (2015), DOI:10.5073/JABFQ.2015.088.033 1 Department of Agricultural Sciences, Alma Mater Studiorum University of Bologna, Bologna, Italy 2 Molecular Phytopathology and Mycotoxin Research Section, Georg-August-University Goettingen, Goettingen, Germany Fusarium proliferatum and fumonisin B1 co-occur with Fusarium species causing Fusarium Head Blight in durum wheat in Italy Barbara Amato1, Katharina Pfohl2, Stefano Tonti1, Paola Nipoti1, Raana Dastjerdi2, Annamaria Pisi1, Petr Karlovsky2, Antonio Prodi1 (Received February 11, 2015) Summary Fusarium Head Blight caused by phytopathogenic Fusarium spp. with Fusarium graminearum as main causal agent is a major disease of durum wheat (Triticum durum Desf.). Mycotoxins in wheat are dominated by trichothecenes B. Fumonisins have only occasionally been reported from wheat; their occurrence was attributed to Fusarium proliferatum and Fusarium verticillioides. We investigated kernels of durum wheat grown in Italy in 2008 - 2010 for colonization with Fusarium spp. and for the content of Fusarium mycotoxins. Fungal biomass was determined using species-specific qPCR and mycotoxins were quantified by HPLC-MS/MS. Fusarium graminearum and Fusarium culmorum were dominating Fusarium species, followed by Fusarium poae, Fusarium tricinctum and Fusarium proliferatum. No Fusarium verticillioides DNA was found. Toxicologically relevant levels of deoxynivalenol and nivalenol but no trichothecenes A were detected. Enniatins, fumonisin B1 and beauvericin were present in grain in all three years. Based on these results and on the evaluation of previous published reports, we hypothesize that low levels of fumonisins commonly occur in wheat grains produced in warm climate; they may remain undetected as long as mycotoxin monitoring programs for wheat do not include fumonisins. The only relevant source of fumonisins in wheat grain appears to be Fusarium proliferatum. Introduction Fusarium head blight (FHB) is a worldwide-occurring cereal disease complex caused by several species of the fungal genus Fusarium that often cause indistinguishable symptoms. The disease leads to the reduction in grain yield, which can reach up to 70 % (PARRY et al., 1995). FHB also affects quality parameters such as the size of kernels, protein content, baking quality of flour and mycotoxin content, making them unsuitable and unsafe for human and animal consumption. The prevailing species in FHB in Italy are F. graminearum Schwabe, F. culmorum (W. G. Smith) Sacc. and recently F. poae (Peck) Wollenweb. (SHAH et al., 2005; PANCALDI et al., 2010). INFANTINO et al. (2012) reported that F. poae was the most prevalent species found in organic bread and durum wheat in Italy. F. graminearum Schwabe [Gibberella zeae (Schwein) Petch] is the predominant FHB pathogen worldwide (XU and NICHOLSON, 2009), producing a range of mycotoxins including type B trichothecenes nivalenol (NIV) and deoxynivalenol (DON). These mycotoxins disrupt cell functions by inhibiting protein synthesis via binding to the ribosome. For this reason, maximum levels of DON as major trichothecene contaminating FHB-affected kernels are regulated in food and feed by European directives EC n. 1881/2006, 126/2007. F. graminearum populations comprise chemotypes that synthesizes either DON or NIV and their acetylated derivatives. F. poae is a pathogen of increasingly importance as a cause of FHB in countries such as Argentina, Canada, Finland, Belgium, Germany, Switzerland, Hungary, Slovakia, Italy, Ireland and the United Kingdom (AUDENAERT et al., 2009; VOGELGSANG et al., 2008; STENGLEIN, 2009; PANCALDI et al., 2010). The prevalence of F. poae in wheat reported in the last years is surprising because the species was believed to be less aggressive than other FHB pathogens (AUDENAERT et al., 2009; XU and NICHOLSON, 2009). Symptoms induced by F. poae are different from those caused by F. graminearum, characterized by small necrotic lesions on the wheat glume (VOGELGSANG et al., 2008). PANCALDI and colleagues (2010) emphasized the importance of F. poae in the disease, showing that the isolation frequency of the species in North Italy increased from 9 % to 23 % from 1998 to 2007. Moreover, in the years where the isolation frequency of F. poae was high, F. graminearum was less often isolated; these observations are consistent with the research by PARRY et al. (1995). Fumonisins are mycotoxins occurring in maize; fumonisins levels in corn are particularly high in Italy due to infection with thermo- philic F. verticillioides. In wheat fumonisins have only occasionally been reported. In wheat plants artificially inoculated with F. pro- liferatum kernel black point symptoms were observed and fumoni- sin B1 was detected (DESJARDINS, 2007), indicating the potential of the fungus as natural contaminant in wheat fields. SHEPHARD et al. (2005) challenged the reports on fumonisins in wheat due to the occurrence of false positive signals when immunoaffinity clean- up, derivatization and fluorescence detection was used in fumonisin detection. More recent work, however, identified fumonisins in wheat unequivocally (e.g. PALACIOS et al., 2011; BUSMAN et al., 2012). The aim of this work was to investigate the occurrence of Fusarium species and their mycotoxins in durum wheat grain in Italy. Materials and methods Experimental design A three years field trial was established in the experimental station of the University of Bologna (Cadriano, 44°33’4.15”N; 11°24’39.02”E), using the durum wheat cultivar Simeto, which is susceptible to FHB. The plot was used for artificial inoculation with F. graminearum and F. culmorum for several years before the sampling. The field was divided in 4 plots (each plot was 3 × 3 m2) for 2008 till 2010, which served as replicates. The first evaluation of FHB symptoms − typical bleaching in the spike − was made three weeks after anthesis on a representative sample of 100 spikes/plot. Disease incidence was determined by counting the number of infected spikes divided by the total number (n=100) of evaluated spikes. number of infected spikes Incidence (DI) = × 100 total number of evaluated spikes Disease severity of every spike was determined by using an evaluation method modified by PURAHONG et al. (2011) that categorizes the bleached spikelet area per spike in percent in eight Fusarium proliferatum and fumonisin B1 in durum wheat 229 severity classes (0 - 2 - 5 -10 - 25 - 50 - 75 - 90 %). Disease severity per plot (DS) was seen as the sum of infected spikes times the disease class divided by the total number of evaluated spikes in percent (including non-diseased spikes). Σ number of infected spikes × disease class Severity (DS) = × 100 total number of evaluated spikes Kernels of every plot were harvested at maturity with a combined harvester. Mycological analysis and species identification Mycological analyses were performed on 25 kernels per plot in 2008 and 2009 and 50 kernels per plot in 2010. Petri dishes, containing PDA (Potato Dextrose Agar, Difco, USA) were set up with 10 kernels each, previously washed in water for 5 minutes, sterilized in 2 % sodium hypochlorite for 5 minutes and rinsed in sterile water. Agar media with kernels were incubated at 25°C for five days in dark. All Fusarium-like colonies were analyzed for their micro- and macroscopical features by optical microscopy (Labor lux 12, Leitz Wetzlar, Germany). Species identification was performed using the synoptic keys by TOUSSOUN and NELSON (1976) and LESLIE and SUMMERELL (2006). To confirm the belonging of a fungus to the species morphologically identified, DNA extraction from mycelium was performed using a CTAB (exadecyl-trimethyl-ammonium bromide) method (PRODI et al., 2011). PCR specific for F. graminearum and F. poae was carried out using primer pairs Fg16 F/R (NICHOLSON et al., 1998) and Fp82 F/R (PARRY and NICHOLSON, 1996), respectively. Amplification was done in a T3 thermocycler (Biometra, Göttingen, Germany) under conditions described in the literature cited above. DNA extraction from wheat flour and real-time PCR analysis The kernels from each plot (2 to 5 kg) were well mixed, 500 g ali- quots were removed and ground. DNA extraction from 1 g of wheat flour was carried out following the upscaled protocol by BRANDFASS and KARLOVSKY (2008). The DNA was dissolved in 200 μl of buffer TE (10 mM Tris HCl, 1 mM EDTA, adjusted to pH 8.0) and diluted 50 times, 1 μl of the dilutions was used as template for real-time PCR. Real-time PCR based on SYBR Green I chemistry was used to eva- luate the level of kernel infection with Fusarium spp. PCR specific for F. culmorum, F. graminearum, F. proliferatum, F. poae, F. verti- cillioides and F. tricinctum (Tab. 1). Serial dilutions of pure fungal DNA as standards were used, where- as the lowest standards were used as limits of quantification, which was 28 ng/g for all Fusarium species. Amplification was done under conditions based on the protocols described by the authors of the primers (Tab. 3) optimized for thermocycler CFX 384 (Biorad, Herkules, CVA, USA) with 384-well plates. DNA standards were quantified by comparative densitometry using known amounts of lambda phage DNA (Biorad Multi-Analyst, Hercules CA, USA). Samples containing DNA below LOQ but with specific melting curve were allocated a value of ½ LOQ. Samples with unspecific melting curves were allocated a value of zero. Mycotoxin analysis Wheat meal was processed, extracted and defatted as described by ADEJUMO et al. (2007). HPLC-MS/MS for the content of DON and NIV was carried out as described below. Non-defatted acetonitrile extracts were kept at -20°C as counter samples. PCR analysis of DNA extracted from the flour carried out four to six years later indicated that kernels could also be contaminated with fumonisins and depsipeptides. Therefore, the counter-samples were re-analyzed for the presence of fumonisin B1, enniatins and beauvericin; because stability of these toxins in acetonitrile extracts over extended periods of time is unknown, the purpose of the analysis was to obtain a lower boundary for their concentration in kernels. Samples for the analysis of beauvericin and enniatins were processed by separating acetonitrile-water supernatants from flour by centrifugation, fol- lowed by drying supernatants in vacuum and redissolving the re- sidues in methanol/water (1:1) without defatting. Samples for the analysis of fumonisin B1 were dried in vacuum, re-extracted with methanol/water (3:1) and cleaned up by SPE on strong anion ex- changer columns (SAX, Agilent, Waldbronn, Germany) according to SHEPHARD et al. (1990). Standards for the quantification of DON and NIV were prepared by spiking uncontaminated wheat flour with pure mycotoxins and processing the spiked samples in the same way as unknown sam- ples (matrix-matched standards). Chromatographic separation on a reverse-phase column, ionization and quantification in multiple reac- tion monitoring mode of a triple quadrupole detector (trichothecenes) and an ion trap detector (the other mycotoxins) was carried out on a system described by RATZINGER et al. (2009), using published mass transitions (RASMUSSEN et al., 2012; ADEJUMO et al., 2007). The limit of quantification (LOQ) was 10 μg/kg for DON, 30 μg/kg for NIV, 10 μg/kg for HT-2 and T-2, 5 μg/kg for DAS, 50 μg/kg for NEO, 3 μg/kg for fumonisin B1, and 14 μg/kg for beauvericin and enniatins. Limits of detection (LOD) values were 3 μg/kg for DON, 10 μg/kg for NIV, 5 μg/kg for HT-2 and T-2, 3 μg/kg for DAS, 20 μg/ kg for NEO, 1 μg/kg for fumonisin B1, and 5 μg/kg for beauvericin and enniatins. Results Fungal colonization and mycotoxin content of kernels In the years 2008 and 2010 the infection pressure was rather high while in 2009 only a limited incidence of FHB was observed (Fig. 1). Disease severity was highest in 2010 and very low in 2009; in spite of the relatively high incidene of FHB in 2008, on the average only 5 % of the spikes were bleached (Fig. 1). Colonization of wheat kernels with F. graminearum as the major causal agent of FHB was investigated by placing 100 to 200 surface- sterilized kernels from each year’s harvest on agar media and count- ing kernels from which F. graminearum isolates were obtained. The isolation frequencies (Tab. 2) matched the ranking of disease severity among years. The biomass of Fusarium spp. in wheat kernels was estimated by species-specific real-time PCR analysis of DNA extracted from representative samples of kernels. The analysis was carried out for the major pathogen of FHB F. graminearum and for the following species: F. culmorum, F. poae, F. proliferatum, F. verticillioides and F. tricinctum (Fig. 2). F. graminearum and F. culmorum were found in high amounts in all three years (except for F. culmorum in 2009). F. poae, F. proliferatum and F. tricinctum occurred in 10 to 100 times lower amounts; no F. verticillioides was found in any sample. Tab. 1: Primers used for species-specific quantification of Fusarium DNA Fungal species Primer name Reference F. culmorum OPT18F /OPT18 R SCHILLING et al., 1996 F. graminearum Fg16 F/R NICHOLSON et al., 1998 F. poae Fp82 F/R PARRY and NICHOLSON, 1996 F. tricinctum Tri1/Tri2 KULIK, 2008 F. proliferatum Fp3-F/Fp4-R JURADO et al., 2006 F. verticillioides VER 1/2 MULE et al., 2004 90 0 230 B. Amato, K. Pfohl, S. Tonti, P. Nipoti, R. Dastjerdi, A. Pisi, P. Karlovsky, A. Prodi Mycotoxins DON and NIV were found in all three years in amounts raising toxicological concerns (Fig. 3); it has to be pointed out, however, that a strong inoculum of trichothecene B producers is likely to be established at the location due to long-year inoculation experiments (see also Discussion). Because PCR analysis revealed the presence of fumonisin and beauvericin producer F. proliferatum and enniatin producer F. tricinctum, the counter samples kept at -20°C were used to determine lower boundaries for concentrations of these toxins in kernels. Fumonisin B1 was found in all samples; four daughter ions generated from the molecular ion possessed the same m/z ratios and occurred in similar relative abundances as daughter ions of authentic fumonisin B1 standard (Fig. 4). Together with the retention time and m/z value of the molecular ion, these results proved convincingly that fumonisin B1 was present in durum wheat kernels. The presence of beauvericin and enniatins A1, B and B1 was supported by their retention times, the molecular ions and three daughter ions for each analyte. Discussion The original goal of our investigation of FHB of durum wheat in Italy was to obtain a deeper insight into the role of F. poae, which attracts increasing attention as a component of FHB complex and a source of highly toxic trichothecenes type A in general (STENGLEIN, 2009) and in Italy in particular (PANCALDI et al., 2010; INFANTINO et al., 2012). Because the area sampled for our analysis was used for inoculation experiments with F. graminearum and F. culmorum in the last years, we expected a strong infection pressure of these patho- gens due to inoculum originating from crop residues. Real-time PCR showed that F. graminearum was the dominant Fusarium species in wheat kernels (Fig. 2). The presence of F. culmorum itself was not surprising. The DNA levels appeared rather high, concerning the climate in Emilia Romagna and the displacement of F. culmorum by F. graminearum in the last decade (YLI-MATTILA, 2010). We speculate that high levels of F. culmorum DNA can be accounted for by the inoculum built after years of inoculation experiments. Colonization of kernels with F. poae was negligible: we found F. poae in very few kernels by agar plating (data not shown) and the content of F. poae DNA in kernels was much lower than the content of F. graminearum DNA (Fig. 2). In line with these results, we have not found any of trichothecenes A typically produced by F. poae (Fig. 3). The presence of F. poae in durum wheat might fluctuate, Fig. 1: Incidence and severity of Fusarium head blight on durum wheat cultivar Simeto grown in a field near Bologna, Italy Tab. 2: Isolation frequency of F. graminearum from naturally infected wheat kernels 2008 2009 2010 F. graminearum 18 2 51 (n=100) (n=100) (n=200) Fig. 2: Fusarium DNA in naturally infected durum wheat kernels Fig. 4: Fumonisin B1 in naturally infected durum wheat kernels Identity of fumonisin B1 was confirmed by comparing authentic standard (blue) with sample extract (red) regarding retention time, m/z of the mother ion and fragmentation spectra Fig. 3: Deoxynivalenol and nivalenol in naturally infected durum wheat kernels Fusarium proliferatum and fumonisin B1 in durum wheat 231 as shown by COVARELLI et al. (2015) who reported that over 20 % of Fusarium isolates from durum wheat ears belonged to F. poae in 2009 but rarely any F. poae was found in the following year. F. tricinctum DNA at moderate levels was found in wheat kernels in two years. Although the species is rarely monitored in FHB, according to published reports it dominated Fusarium microflora of wheat in Syria (ALKADRI et al., 2013) and was one of the dominant species in barley in (NIELSEN et al., 2014). qPCR analysis revealed the presence of DNA of F. proliferatum in all years. In year 2009, colonization of wheat kernels with F. proliferatum was comparable to the colonization with F. culmorum (Fig. 2). F. proliferatum produces fumonisins, which are proven carcinogens in rodents and suspected carcinogens in humans. Contamination of wheat with fumonisins is therefore highly relevant for food safety. Fumonisins regularly occur in maize grown in warm climate; major fumonisin producers in maize are F. verticillioides and F. proliferatum (FANDOHAN et al., 2003) with F. verticillioides often reported to dominate (REYES-VELAZQUEZ et al., 2011). While trichothecene content in wheat has been documented in hundreds of publications, fumonisins were only occasionally detected in wheat (DESJARDINS et al., 2007; PALACIOS et al., 2011; BUSMAN et al., 2012). Founding DNA of F. proliferatum in the total DNA extracted from wheat kernels motivated us to extend the analysis of mycotoxin content to fumonisins and beauvericin, which are both produced by F. proliferatum. We also analyzed the content of enniatins, which are produced by many Fusarium spp. including F. tricinctum which we detected in wheat kernels in two years (Fig. 2). Fumonisins were detectable in wheat kernels in all three years, though the biomass of F. proliferatum was hundred times lower than the biomass of F. graminearum. In analogy to maize, the origin of fumonisins in wheat is generally attributed to F. proliferatum and F. verticillioides; the results re- ported by INFANTINO et al. (2012) indicated that F. verticillioides could be the major source of fumonisins in wheat in Italy. Our durum wheat samples were consistently contaminated with fumonisins but we have not found any DNA of F. verticillioides. F. proliferatum, which was found in the kernels consistently in all three years, appeared to be the only source of fumonisin B1. F. verticillioides is ubiquitous in Italian maize fields; inoculum in form of airborn conidia likely reached wheat fields, too. We are confident that our qPCR analysis reliably distinguished between F. proliferatum and F. verticillioides. The primers targeted intergenic spacers of rRNA genes, which were shown to be taxonomically informative for the differentiation between the two species (VISENTIN et al., 2009), and they have been extensively tested before (NUTZ et al., 2011). Even if F. verticillioides can infect wheat seeds, as reported by INFANTINO et al. (2012), fumonisins in wheat do not seem to originate from F. verticillioides because DESJARDINS et al. (2007) and BUSMAN et al. (2012) reported that F. verticillioides artificially inoculated into wheat did not produce any fumonisins, even thought the isolates produced large amounts of fumonisins in vitro and in inoculated maize plants. The DON and NIV levels in kernels reached toxicologically relevant levels in all three years. In 2008 and 2010 the amount of DON exceeded by a large margin the legal limit of 1,750 μg/kg for unprocessed durum wheat established by European Regulations EC n. 1881/2006 and EC n. 1126/2007. The high levels of DON can likely be accounted for by inoculation experiments carried out in the field with F. graminearum in last years, which cause build- up of inoculum on crop residues. Which Fusarium species was the source of DON and NIV in our samples? We found a significant positive correlation between the concentration of NIV and the amount of DNA of F. graminearum in kernels. DON concentration and F. graminearum DNA were positively correlated, too, but the significance of the correlation was not statistically supported (data not shown). The amount of F. culmorum DNA in kernels has not correlated with NIV nor DON. We therefore assume that the dominant source of trichothecenes in kernels was F. graminearum, which is in line with the view that F. graminearum is the major causal agent of FHB in Europe (YLI-MATTILA, 2010). The amount of DON in grain exceeded the amount of NIV by a factor of 4 to 10, which is in line with the observation that F. graminearum chemotypes producing DON and its derivatives dominate over NIV producers in wheat in North Italy (PRODI et al., 2009; PRODI et al., 2011). High competitiveness of F. graminearum was reported by XU and NICHOLSON (2009) in co-inoculations in vitro trials in controlled environmental conditions. Wheat kernels in our samples were con- sistently colonized with F. proliferatum though they contained high levels of F. graminearum DNA. Co-inoculation experiments with F. proliferatum have not been conducted in wheat so far. To elucidate the interaction of F. proliferatum and other minor members of the Fusarium complex on wheat kernels with F. graminearum, we intend to investigate individual ears and kernels for the presence of Fusarium species. Acknowledgements The authors would like to thanks Stefano Borsari for his help in the preparation of the materials for the field experiments and Dr. Davide Pancaldi for his suggestions and comments as well as Heike Rollwage for technical assistance. References Adejumo, T.o., HeTTwer, u., KArlovsKy, P., 2007: Survey of maize from south-western Nigeria for zearalenone, alpha- and beta-zearalenols, Tab. 3: Lower boundary for the content of fumonisin B1, beauvericin and enniatins The content of fumonisins and depsipeptides was determined in samples that were stored for 4 to 6 years in acetonitrile-water at -20°C. Because stability of mycotoxins under these conditions is unknown, we regard the values as lower boundaries for mycotoxin content in freshly harvested kernels. Mean of four plots per year ± standard deviation is shown except for enniatin B, for which all four values are listed. Limit of detection (LOQ) was 3 ng/g for fumonisin B1 and 14 ng/g for the other mycotoxins. “n.d.”: signal not detectable; “