International Journal of Food Microbiology 149 (2011) 226–235
Contents lists available at ScienceDirect
International Journal of Food Microbiology
j ourna l homepage: www.e lsev ie r.com/ locate / i j foodmicro
Quantification of ochratoxin A-producing molds in food products by SYBR Green andTaqMan real-time PCR methods
Alicia Rodríguez a, Mar Rodríguez a, M. Isabel Luque a, Annemarie F. Justesen b, Juan J. Córdoba a,⁎a Food Hygiene and Safety, Faculty of Veterinary Science, University of Extremadura, Avda. de la Universidad, s/n. 10071-Cáceres, Spainb Department of Integrated Pest Management, Faculty of Agricultural Sciences, University of Aarhus, Denmark
Ochratoxin A (OTA) is a mycotoxin synthesized by a variety of different fungi, most of them from the generaPenicillium and Aspergillus. Early detection and quantification of OTA producing species is crucial to improvefood safety. In the present work, two protocols of real-time qPCR based on SYBR Green and TaqMan weredeveloped, and their sensitivity and specificity were evaluated. Primers and probes were designed from thenon-ribosomal peptide synthetase (otanpsPN) gene involved in OTA biosynthesis. Seventy five mold strainsrepresenting OTA producers and non-producers of different species, usually reported in food products, wereused as references. All strains were tested for OTA production by mycellar electrokinetic capillaryelectrophoresis (MECE) and high-pressure liquid chromatography-mass spectrometry (HPLC-MS). Theability of the optimized qPCR protocols to quantify OTA-producing molds was evaluated in differentartificially inoculated foods. A good linear correlation was obtained over the range 1 x 104 to 10 conidia/g perreaction for all qPCR assays in the different food matrices (cooked and cured products and fruits). Thedetection limit in all inoculated foods ranged between 1 and 10 conidia/g for SYBR Green assay and TaqMan.No significant differences were found between the Ct values obtained from pure mold DNA and pure moldDNA mixed with food DNA. The ability of the designed qPCR methods to quantify two known conidialsuspensions inoculated on several foods was evaluated. The amount of conidia assessed by both qPCRmethods was close to the inoculated amount for most foods and indicates that the described procedure holdspotential for use for the detection and quantification of OTA producing molds in foods.
Mycotoxin producing molds are among the main sources ofcontamination in foods (Sweeney and Dobson, 1998), especially inripened foods. The environmental conditions found in meats andcheeses throughout the ripening process favor growth of a moldpopulation composed mainly of Penicillium and Aspergillus species(Kure et al., 2004; Núñez et al., 1996). In addition, molds of thesegenera are also contaminants in other foods such as grapes, cereals,coffee, raisins, spices and wine, (Abarca et al., 2003; Ayoud et al.,2010; Bellí et al., 2004; Gil-Serna et al., 2009, 2011; MikuŠovà et al.,2010; Thirumala-Devi et al., 2001).
Several species of both genera isolated from these products havebeen reported as ochratoxin A (OTA) producers (Dachoupakan et al.,2009; López-Mendoza et al., 2009; Lund and Frisvad, 2003; Mulé et al.,2006; Sartori et al., 2006). OTA is a nephrotoxic polyketide mycotoxincoupled to the amino acid phenylalanine synthesized by a variety ofdifferent fungi, most of them from Aspergillus genus, in particular, A.
ochraceus, A. carbonarius or A. niger (Karolewiez and Geisen, 2005).Furthermore, two new species from Aspergillus section Circumdati suchas A. westerdijkiae and A. steynii which have recently been split fromA. ochraceus have been reported to be stronger OTA-producing moldsthan A. ochraceus (Gil-Serna et al., 2011). Aspergillus species areresponsible for the occurrence of OTA in products such as coffee, raisins,grape juice, spices or wines (Gil-Serna et al., 2011; González-Salgadoet al., 2009; Sartori et al., 2006). Furthermore, two species of Penicilliumgenus, P. verrucosum and P. nordicum are both known to produce OTA(Bogs et al., 2006; Lund and Frisvad, 2003).
Early detection of OTA producing species is crucial in order toprevent OTA from entering the food chain, to improve food safety andto protect consumers from hazardous mycotoxins. For this purposeDNA-based techniques are good alternatives to traditional identifica-tion techniques, as they are rapid, sensitive, specific and allowaccurate identification of fungal species (Gil-Serna et al., 2009; Selmaet al., 2008). Real-time quantitative PCR (qPCR) moreover provides atool for accurate and sensitive quantification of target DNA (Muléet al., 2006; González-Salgado et al., 2009). For the development ofspecific qPCR assays, it is essential to use sensitive methods tomeasure the OTA production from reference mold strains. Mycellarelectrokinetic capillary electrophoresis (MECE) and high-pressure
227A. Rodríguez et al. / International Journal of Food Microbiology 149 (2011) 226–235
liquid chromatography-mass spectrometry (HPLC-MS) have beenreported as sensitive methods to detect OTA production (Martín et al.,2004; Sosa et al., 2002).
Real-time qPCR enables detection and quantification of amplifiedPCR product by continuous monitoring of the fluorescence from non-specific dyes, such as SYBR Green, which can also give a signal forprimer-dimers and non-specific amplified products (Kubista et al.,2006) or from a sequence specific hydrolysis probe (TaqMan).Recently, several qPCR methods for detection of OTA producingmold species from different food products have been reported, such asfor A. carbonarius in wine grapes (Selma et al., 2008), P. verrucosum incereals (Schmidt-Heydt et al., 2007) or P. nordicum in meat (Bogset al., 2006). However, no qPCR assay has been designed for detectionof OTA producers regardless of the mold species. The correct choice oftarget sequence for design of primers is essential for the developmentof new qPCR protocols for detection and quantification of OTAproducing strains of different mold species and genera. Although, theOTA biosynthesis pathway has yet not been completely elucidated, anumber of putative pathways have been proposed (Harris andMantle,2001; Huff and Hamilton, 1979). According to the molecular structureof OTA it is clear that a number of enzymatic reactions are likely to berequired for its biosynthesis: a polyketide synthase for the synthesisof the polyketide dihydroisocoumarin, a cyclase, a chloroperoxidaseor halogenase, an esterase and a peptide synthetase for ligation of thephenylalanine to the dihydroisocoumarin (Gallo et al., 2009). Thenon-ribosomal peptide synthetase (otanpsPN) gene, responsible forthe linkage of the phenylalanine moiety to the polyketide, has beenreported as a key enzyme of the OTA biosynthesis gene cluster fromPenicillium and Aspergillus species (Abbas et al., 2009; Bogs et al.,2006; Harris and Mantle, 2001; Karolewiez and Geisen, 2005). Thisgene has successfully been used for the detection of OTA-producingmolds by conventional PCR (Bogs et al., 2006; Geisen et al., 2006).
The efficiency of PCR quantification methods can be seriouslyaffected by the presence of inhibitors in the food matrix. According toWilson (1997), they may interfere at several steps in the quantifica-tion procedures e.g. cell lysis necessary for extraction of DNA, nucleicacid degradation or capture and also by inhibiting the polymeraseactivity for amplification of target DNA. Inhibition may be expressedas false negatives or reduced sensitivity. Organic compounds such aspolyphenols and polysaccharides from vegetables and ripened foodscan also lead to errors in the specific detection of the OTA producers(Monnet et al., 2006; Mulé et al., 2006). Therefore, direct testing insuch foods is required to evaluate the potential application of qPCRbased methods for detection and quantification of molds. The aim ofthe present work was to develop sensitive and specific qPCR methodsfor detection and quantification of OTA-producing molds in foods.
2. Material and methods
2.1. Fungal strains
Seventy-five mold strains, belonging to different species, wereobtained from the Spanish Type Culture Collection (CECT), theCentraalbureau voor Schimmelcultures in The Netherlands (CBS),the Type Culture Collection of the Department of Biotechnology fromthe Technical University of Denmark (IBT), the Culture Collection ofFood Hygiene and Safety from University of Extremadura and strainskindly supplied by Dr. Covadonga Vázquez (University Complutenseof Madrid, Spain). Only 12 of them were known as OTA-producers,whereas no information on OTA production was available for theremaining strains, as indicated in Table 1.
2.2. OTA production
Production of OTA was tested after growing the mold strains inPotato Dextrose Agar (Sharlau Chemie S.A., Spain) for 15 days at 25 °C.
OTA production was analyzed by MECE, based on the spectrum ofabsorbance between 190 and 600 nm (Martín et al., 2004), and also byHPLC-MS, obtaining the full MS spectra after atmospheric pressurechemical ionization (Sosa et al., 2002).
2.3. DNA extraction
All mold strains were 3-point inoculated on Malt Extract Agar (2%malt extract, 2% glucose, 0.1% peptone, and 2% agar) and incubated for4 days at 25 °C. Grown mycelium was scraped off the agar and about50 mg of isolated mycelium from each strain were used for genomicDNA extraction following the method described by Sánchez et al.(2008). DNA concentration was quantified spectrophotometricallywith a Biophotometer Eppendorf (Eppendorf AG, Hamburg, Germany)and1.0 ng of DNAextracted fromproducing andnon producing strainswas used for qPCR assays.
2.4. Primers and probes design
The primers otanps_ for and otanps_rev targeting the otanpsPNgene and previously used in a conventional PCR described by Bogs etal. (2006)was applied on genomic DNA from both OTA-producing andnon-producing strains. An amplicon of 750 bp was obtained frommost OTA-producing strains of Penicillium, Aspergillus, and Emericellaspecies. This amplicon was purified, sequenced and analyzed. Then, aspecific primer pair F/R-npstr (Table 2) amplifying an amplicon of117 bp was designed from conserved regions using Primer Expresssoftware (Applied Biosystems, Foster City, CA, USA). For testing thespecificity of the obtained amplicons, they were sequenced andaligned with the published sequence of the otanpsPN gene (GenBankaccession no. AY557343) using ClustalW2 (http://www.ebi.ac.uk/Tools/msa/clustalw2/) as indicated in Table 2.
In addition, the NPSprobe TaqMan probe was designed for theTaqMan assay (Table 2). This probe was labeled at the 5′ end with thereporter cyanine-5 (Cy5) and at the 3′ end with the quencher BlackHole Quencher 2 (BHQ2) (Sigma Aldrich, Madrid, Spain).
2.5. qPCR reactions
The Applied Biosystems 7500 Fast Real-Time PCR system (AppliedBiosystems) was used for qPCR amplification and detection. qPCRwasprepared in triplicates of 20 or 25 μL reaction mixture in MicroAmpoptical 96-well reaction plates and sealedwith optical adhesive covers(Applied Biosystems). Three replicates of a control sample withoutDNA template were also included in the runs.
2.5.1. SYBR Green qPCRThe primer pair F/R-npstr was first evaluated in a SYBR Green
protocol. To optimize the amount of primers and template DNA in thereactions, one OTA-producing strain P. viridicatum CECT 2320 wasused. The tested concentrations ranged from 700 to 200 nM forprimers and from 10 ng to 0.1 pg for template DNA.
The optimized SYBR Green protocol was carried out in a finalvolume of 25 μL, containing 1.0 ng of template DNA, 12.5 μL of 2×SYBR® Premix Ex TaqTM (Takara), 0.5 μL of 50× ROXTM Reference Dye(Takara) and 400 nM of both F-npstr and R-npstr primers.
The following thermal cycling conditions were used by the SYBRGreen method: a single step of 10 min at 95 °C, 40 cycles of 95 °C for15 s and 60 °C for 1 min. After the final PCR cycle, melting curveanalysis of the PCR products was performed by heating to 60–95 °Cand continuous measurement of the fluorescence to verify the PCRproduct. Threshold cycle (Ct) value, corresponding to the PCR cyclenumber at which fluorescence was detected above threshold, wascalculated by the 7500 Fast System SDS software (Applied Biosystems).Allmold strains used in this studywere testedby triplicatewith the abovemethod. The size of PCR products was verified by electrophoresis in 2.5%
Table 1OTA production and qPCR data with the specific primers (F-npstr and R-npstr) and NPSprobe of the reference mold strains.
228 A. Rodríguez et al. / International Journal of Food Microbiology 149 (2011) 226–235
No Template Control (NTC) Ct value was established in 31 and 33 for SYBR Green and TaqMan assay, respectively.All OTA producing molds strains are shaded.a Data represent the mean threshold cycle (Ct)±standard deviation (SD) of the 3 independent experiments each consisting of triplicate samples.b Tm: melting temperature.c CBS: Centraalbureau voor Schimmelcultures (The Netherlands).d OTA concentration (μg/l) detected by MECE or HPLC-MS.e CECT: Spanish Type Culture Collection.f ND: OTA production is not detected by MECE or HPLC-MS.g IBT: Type Culture Collection of the Department of Biotechnology (Technical University of Denmark).h Pc: strains isolated from dry-cured ham belonging to the Culture Collection of Food Hygiene from University of Extremadura.i Strains kindly supplied by Dr Covadonga Vázquez (University Complutense of Madrid, Spain).
Table 1 (continued)
229A. Rodríguez et al. / International Journal of Food Microbiology 149 (2011) 226–235
agarose gels. These gels were stained with ethidium bromide andvisualized with UV transillumination. The PCR product was also purifiedand sequenced to check if its sequence was from the otanpsPN gene.
2.5.2. TaqMan qPCR conditionsThe F/R-npstr primer pair and the NPSprobewere assayed with the
OTA producing strain P. viridicatum CECT 2320 for the TaqMan-basedPCR. To optimize the reactions, several concentrations ranging from400 to 125 nM for primers, 500 nM to 100 nM for probe, and 10 ng to0.1 pg for template DNA were assayed. The reaction mixture for thisassay consisted of 10.0 μL of Premix Ex TaqTM (Takara), 0.4 μL of ROX of50× ROXTM Reference Dye (Takara), 400 nM of each F-npstr andR-npstr primers, 500 nM of the NPSprobe, and 4 μL of template DNAwith a concentration of 0.25 ng/μL in a final volume of 20 μL. Thethermal cycling conditions included an incubation of 2 min at 50 °C toallow the uracil-N-glycosylase (UNG) enzyme was activated, anincubation step for 10 min at 95 °C to denature the UNG enzyme andactivate AmpliTaq Gold polymerase, 40 cycles at 95 °C for 15 s and60 °C for 1 min. Ct determinations were automatically performed bythe instrument using default parameters. All mold strains used in thisstudy were tested in triplicate. The size of PCR products was verifiedby electrophoresis in 2.5% agarose gels. These gels were stained withethidium bromide and visualized with UV transillumination. The PCRproduct was also purified and sequenced to verify if its sequence wasfrom the otanpsPN gene.
2.6. Sensitivity and specificity of qPCR reactions
P. verrucosum CBS 323.92, one OTA-producing strain, was used toevaluate the sensitivity of the designedmethods. For this purpose, 10-fold DNA dilutions of the above strain ranging from 10 ng to 0.001 pgwere prepared. For qPCR amplification, 3 replicates of each DNAdilution were assayed per run. Standard deviations were calculatedfor each treatment between technical replicates. The detection limit ofthe producing strain was defined as the lowest DNA concentrationdetected in all reactions. The criteria considered for reliability of thedesigned methods were the correlation coefficient and the amplifi-cation efficiency calculated from the formula E=10−1/S−1 (S beingthe slope of the linear fit).
The specificity and sensitivity of SYBRGreen and TaqMan assayswastested on genomic DNA mixtures of a non OTA-producing strain(P. aurantiogriseum, CBS 112021) and an OTA-producing strain(P. viridicatum, CECT 2320). Ten-fold DNA dilutions of the OTA-producing strain ranging from 10 ng to 0.001 pg were mixed with10 ng of DNA from the non-producing strain. In order to evaluate thespecificity of the amplicons, the melting temperature (Tm) was
automatically calculated in the SYBR Green assay and compared withthat deduced from the sequence of the expected fragment.
Β-tubulin partial sequences of several new OTA-producing species aswell as ITS1–5.8S–ITS2 region, were obtained in order to confirm theiridentification. PCR reactions forβ-tubulinwere performed using primersBt2a and Bt2b (Glass and Donaldson, 1995). The amplification programused was: 1 cycle of 5 min at 94 °C, 32 cycles of 1 min at 94 °C, 1 min at68 °C and 1 min at 72 °C and finally 1 cycle of 5 min at 72 °C. PCRreactions for ITS1–5.8S–ITS2 region were performed using primers ITS1and ITS4 (White et al., 1990). The amplification program used was:1 cycle of 5 min at 94 °C, 40 cycles of 1 min at 94 °C, 1 min at 50 °C and2 minat72 °Candfinally1 cycleof 5 minat 72 °C.Amplificationproductswere purified and sequenced as described in Section 2.4.
2.7. Sensitivity of qPCR for OTA producing molds on artificially inoculatedfood matrices
The sensitivity of the optimized qPCR methods was assayed withDNA extracted from three types of food products, each inoculatedwith spores of different OTA-producing strains. Cooked ham wasinoculated with P. aurantiogriseum CECT 2264, ripened cheese withP. nordicum CBS110769, andgrapewithA. ochraceusCBS589.68. To testthe specificity of both qPCRmethods when applied on food, three typesof foods were inoculated with an OTA producer A. westerdijkiae (CECT2948) and non-producer A. westerdijkiae (3.38). For this, the spores ofthe above producing and non producing strains were harvested byflooding 3 plates (20 days old) of Malt Extract Agar with 5 mL of sterilenanopure water containing 10% glycerol (Scharlau Chemie S.A., Spain),and rubbing the surface with a glass rod. The conidial suspension wasfiltered throughWhatmanpaper No. 1, diluted in sterile nanopurewaterand quantified by microscopy, using a Neubauer counting chamber.Aliquots of 1 mLof 102, 103, 104, 105, and106 conidiamL−1 of eachmoldstrainwere immediately used to spike food samples, to concentrationsof1, 10, 102, 103, and, 104 conidia per gram of food.
For DNA extraction, 5 g of spiked food samples were homogenizedwith 10 mL of Tris–HCl buffer (pH 8.0) in a filter bag BagPage(Interscience, Paris, France) using a pulsifier equipment (Microgenbioproducts, Surrey, UK). The filtrate obtained within each bag wastransferred to a sterile tube, and centrifuged at 13,000 rpm for 10 min.Pellets were resuspended in 100 μL of sterile nanopure water, boiled(95 °C for 15 min) to release the DNA, and cooled on ice for 10 min.Next, 500 μL CTAB buffer (5 g D-sorbitol, 2 g N-lauroylsarcosine, 1.6 g/L CTAB, 1.4 M NaCl, 20 mM Na2EDTA, 2 g PVPP, 0.1 M Tris–HCl, pH8.0) was added together with 10 μL of a proteinase K solution (10 mg/mL), 10 μL of a lyticase solution (1 mg/mL) (for cooked ham andgrape) and 5 μL of a 2-mercaptoethanol (only for grape) (Sigma-
230 A. Rodríguez et al. / International Journal of Food Microbiology 149 (2011) 226–235
Aldrich Química, S.A., Madrid, Spain) before incubation at 65 °C for1 h. Samples were centrifuged at 13,000 rpm for 5 min, and thesupernatant was transferred to a new tube with 500 μL chloroform,vortexed, and centrifuged at 13,000 rpm for 20 min. The upper layerwas transferred to a new tube and 10 μL RNase solution (10 mg/mL)was added before incubation at 37 °C for 1 h. An equal volume ofchloroform was then added, vortexed, and centrifuged at 13,000 rpmfor 5 min. Finally, the aqueous phase was processed according to theEZNA Fungal DNAMini Kit (Omega bio-teck, Doraville, USA) protocol,starting from DNA precipitation by adding 500 μL of isopropanol (step4 protocol B). In the final step, DNA was eluted in 100 μL of elutionbuffer pre-warmed to 65 °C and kept at −20 °C until used as templatefor PCR amplification. All inoculations and extractions were performedby triplicate for each food. In addition, triplicates of non-inoculatednegative control were included in each experiment.
For qPCR amplification, 3 replicates of 5 μL DNA extracted from thespiked foods and negative controls without conidia were assayed perrun. Standard curves were generated for each group of food productsand the efficiencies for each standard curve were calculated.
To directly estimate the load of the inoculated mold, the sampleswere diluted in 0.1% peptone water and spread-plated on PotatoDextrose Agar (Sharlau Chemie S.A., Spain) and incubated at 25 °C for4 days. The natural fungal contamination of the sampleswas lower than1 log cfu/g, and aftermorphological characterization of the isolates noneP. aurantiogriseum, P. nordicum, A. ochraceus spp and A. westerdijkiaewere found. In addition,DNAof this fungal isolates yieldnegative resultsin the developed qPCR with SYBR Green.
In order to confirm the lack of PCR inhibitors in the food matrices,five independent genomic DNA extractions from three OTA-producingmolds P. aurantiogriseum CECT 2264, P. nordicum CBS 110769 andA. ochraceus CBS 589.68 were spiked in equal amounts (0.05 ng) tototal DNA extracted from different amounts (10, 5 and 1 g) of non-inoculated cooked ham, ripened cheese and grape, respectively.Amplification plots were compared with those obtained by amplifyingthe same pure strains diluted in deionized water.
2.8. Quantification of conidia in artificially inoculated foods by qPCR
Fungal spores of different OTA producing strains were inoculated innine different foods, including cooked meat products (mortadella,cooked turkey breast and cooked ham), ripened foods (dry-cured ham,dry-fermented sausage (“salchichón”) and ripened cheese) and fruits(grape, plumb and pear). Inoculations were carried out as follows:cookedmeat productswith P. aurantiogriseumCECT 2264, ripened foodswith P. nordicum CBS 110769, and fruits with A.ochraceus CBS 589.68.Three independent tests were run at two levels of fungal spores: low(3×10 conidia/g) and high (3×103 conidia/g). DNA-extraction wascarried out as described in Section 2.7 for artificially inoculated foodmatrices. qPCRwere carried out using 2.5 μL of DNA in triplicate. The Ctvalues for inoculated samples were extrapolated from standard curvesobtained from artificially inoculated foods. Triplicates of non-inoculatednegative controls were included in each experiment.
Table 2Nucleotide sequence of primers and probe used for either SYBR Green or TaqMan RealTime PCR assays.
Primerpair
Primername
Nucleotide sequences(5′-3′)
Productsize
aPosition
F/R-npstr F-npstr GCCGCCCTCTGTCATTCCAAG 117 bp 5090R-npstr GCCATCTCCAAACTCAAGCGTG 5185
NPSprobe [Cy5]-CGGCCGACCTCGGGAGAGA[BHQ2] 5145
a Positions are in accordance with the published sequences of otanpsPN gene ofP. nordicum (GeneBank accession no. AY557343).
2.9. Statistical analysis
All the statistical analyses were performed with the SPSS v.15.0.Oneway analysis of variance (ANOVA)was carried out to determinatesignificant differences within and between groups. Tukey's test wasapplied to compare the mean values. Statistical significance was set atP≤0.05.
3. Results
3.1. Detection of OTA by MECE and HPLC-MS
MECE and HPLC-MS analyses confirmed OTA production in all 12strains identified as OTA producers by the Culture Collections (Table 1).In addition, 13 strains with no information on OTA production in theCulture Collections producedOTA,whereas the 50 remaining strains didnot produce detectable amounts.
The concentration levels of OTA produced by the OTA-producingstrains detected in PDA medium ranged from 1 to 100 μg/l.
3.2. Analysis of OTAnpsPN gene sequences of OTA-producing species
The DNA fragment of 750 bp from OTAnpsPN gene of the differentspecies targeted was amplified, and sequenced. These sequences wereanalyzed anddeposited inGenBankmatchingwithNPSgenes (GenBankaccession nos. JN097797; JN097798; JN097799; JN097800, JN097801,JN097802, JN097803, JN097804, JN097805, JN097806, JN097807,JN097808). All these sequences were compared with the publishedsequence of OTAnpsPN gene of P. nordicum (GenBank accession no.AY557343) showed a similarity greater than 98% in all cases.
3.3. Optimization of qPCR conditions
The F/R-npstr primer pair gave one PCR product of the expectedsize with a Tm of 84.6±0.3 °C when using the SYBR Green method. Inaddition, no considerable nonspecific amplification, including primer-dimers, was observed.
The best primers and probe concentrations giving the lowest Ctvalue with an adequate fluorescence for a given target concentrationwere selected for further analyses. Optimal F-npstr and R-npstrconcentrations used for the SYBR Green reactions were 400 nM inboth cases. For the TaqMan assays, the lowest Ct value was obtainedwith 400 nM of each primer and 450 nM of probe.
3.4. Evaluation of reference strains by the optimized SYBR Green andTaqMan qPCR protocols
All OTA-producing reference strains detected by MECE and HPLC-MS analysis, including the reference strains belonging to Penicillium,Aspergillus and Emericella, showed Tm values ranging from 84.0 to84.9 °C in the SYBR Green method, while non-producing strains fromthe above genera ranged from 65.6 to 81.2 °C (Table 1). This wasconfirmed with agarose gel electrophoresis which showed a 117 bpproduct from all OTA-producers only (data not shown). OTA pro-ducers showed Ct values lower than 32 and 30 in SYBR Green andTaqManmethods, respectively. Itwas ensured that the sameamounts oftemplate DNA was present in the reactions as indicated in Section 2.3.Furthermore, significant differences (pb0.01) were found between theCt values obtained from OTA producing and non-producing referencestrains for both methods.
3.5. Standard curves, sensitivity, and detection limits of qPCR usingSYBR-Green and TaqMan systems
The sensitivity of the qPCR assays was evaluated by testing 10-folddilution series from10 ng to 0.001 pgDNA fromP. verrucosum strain CBS
Fig. 2. Standard curves showing the log10 DNA amount vs. threshold cycle (Ct) values ofSYBR Green (a) and TaqMan (b) Real-Time PCR methods for different amounts of DNAfrom of the OTA-producer P.verrucosum CBS 323.92.
231A. Rodríguez et al. / International Journal of Food Microbiology 149 (2011) 226–235
323.92. The Tmand Ct values obtained from theDNAamounts tested foreach qPCRmethod and the analysis in agarose gel electrophoresis of theqPCR products are shown in Fig. 1. The detection limits were 0.01 pg forboth SYBR Green and TaqMan assays.
Standard curves relating the Ct values and the amount ofP. verrucosum CBS 323.92 purified genomic DNA were generated foreach optimizedmethod (Fig. 2). A good linear relationship between theincreasing Ct values and the target DNA was observed over the range10 ng to 0.01 pg for both developed qPCR methods. The slopes of thelinear regression curves were −3.34 for SYBR Green assay, and −3.20for the TaqMan assay (Fig. 2). Therefore, the efficiencies of the SYBRGreen and the TaqMan assays were 99.1 and 105.4%, respectively.Furthermore, the correlation coefficients were 0.99 in both cases.
3.6. Evaluation of primers and probe specificity for SYBR Green andTaqman assays
In the SYBR Green assay, amplification with Ct values lower than32 was obtained with more than 0.1 pg DNA from the producingstrain in presence of 10 ng of DNA from a non-producing strain(Fig. 3). To evaluate the specificity of the amplicons, the melting pointwas calculated automatically, and the mean values obtained withmore than 0.1 pg DNA were 84.4±0.2 °C. Additionally, a specific117 bp PCR product was detected by agarose gel electrophoresis(Fig. 3). However, when the quantity of DNA from the producingstrain was below 0.1 pg, different Tm values were obtained rangingfrom 76.9 to 77.8 °C, and a PCR product was not detected by agarosegel electrophoresis (Fig. 3).
For the TaqMan assay, Ct values lower than 35 were obtained withat least 0.1 pg DNA from an OTA-producing strain mixed with 10 ngDNA from a non-producing strain (Fig. 3). A specific 117 bp PCRproduct was detected by agarose gel electrophoresis.
3.7. Sensitivity of the qPCR assay on artificially inoculated food matrices
The ability of the optimized qPCR protocols to quantify OTA-producing molds was evaluated in different artificially spiked foods.Standard curves using DNA extracted from inoculated foods weregenerated for each food matrix. The slopes of the linear regressionequations in the SYBR Green assays for cooked ham, ripened cheeseand grape were respectively −3.32, −3.30 and −3.63, while withTaqMan assays the slope values were −3.20, −3.50 and −3.64,respectively. A good linear correlation (R2) was also obtained over therange 1×104 to 10 conidia/g per reaction for all the food matrices
Fig. 1. Sensitivities of SYBR Green and TaqMan Real-Time PCR assays for the detection fro
(Table 3). The efficiencies values ranged from 88 to 105% (Table 3).The detection limit in all inoculated foods was 10 conidia/g for theTaqManaswell as for the SYBRGreen assay thesewere applied on fruits.In cooked and ripened cheese, the detection limitwas1 conidia/g for theSYBR Green assay.
The robustness of qPCR methods was tested in three food matriceswhich were inoculated with a producing and a non-producingA. westerdijkiae strain. Standard curves were constructed for eachmatrix. The efficiencies values obtained in all inoculated food with theproducing A. westerdijkiae strains were similar to the other producingstrains used in the sensitivity assay from inoculated foods (Table 3). In
on of OTA producer strain P. verrucosum CBS 323.92 (ng)
0.01 0.001 0.0001 0.00001 0.000001 Negative control
Fig. 3. Specificities of SYBR Green and TaqMan Real-Time PCR assays for detection from 1 ng to 0.01 pg of DNA of an OTA-producing mold (P. viridicatum CECT 2320) with thepresence of 10 ng of DNA of a non-producing strain (P. aurantiogriseum CBS 112021).
232 A. Rodríguez et al. / International Journal of Food Microbiology 149 (2011) 226–235
addition, no amplificationwas observedwhen the foodswere inoculatedwith non-producing A. westerdijkiae strain (Ct values around of 40).
The absence of PCR inhibitors was studied in all of the inoculatedfoods by the comparisonof the amplification plots of pureOTAproducerDNA and pure OTA producer spiked with non-inoculated food sampleDNA extracted from 10 to 1 g of sample. No significant differences werefound between the Ct values obtained from OTA producing mold DNA,pure and pure spiked with food DNA (Table 4).
The ability of the designed qPCR methods to quantify two knownamounts of conidia added to several foodmatrices is shown in Table 5.The amount of conidia assessed by both qPCR methods was close tothe inoculated amount in most food matrices when the concentrationwas 3×103conidia/g. No significant effect was observed due to thedifferent food matrices and qPCR methodology used (pb0.05). At alower concentration of conidia (30 conidia/g) a larger variation wasseen (Table 5). Nevertheless, no significant differences were observeddue to different food matrices or qPCR method employed. Thestandard deviation in the SYBR Green assays was from 0.19 to 0.91and in the TaqMan assays were from 0.21 to 0.68.
4. Discussion
In the present work, the otanpsPN gene was target for thedevelopment of specific qPCR assays for detecting OTA-producing
Table 3Efficiencies of amplification and R2 obtained from standard curves of conidia from OTAproducing molds in artificially inoculated food by qPCR.
Food sample OTA producingspecies
qPCRmethod
Efficiency(%)
R2
Cooked ham P. aurantiogriseum CECT 2264 SYBR Green a100±0.9 0.98TaqMan 105 ±1.2 0.99
A. westerdijkiae CECT 2948 SYBR Green 101±0.7 0.98TaqMan 104±1.2 0.98
Ripened cheese P. nordicum CBS 110769 SYBR Green 100±0.7 0.99TaqMan 93±0.3 0.98
A. westerdijkiae CECT 2948 SYBR Green 99±1.3 0.98TaqMan 95±0.5 0.98
Grape A. ochraceus CBS 589.68 SYBR Green 89±1.4 0.98TaqMan 88±2.2 0.99
A. westerdijkiae CECT 2948 SYBR Green 89±1.4 0.98TaqMan 89±1.2 0.99
a Data represent the mean efficiency±standard deviation of the 3 independentassays each consisting of triplicate samples.
molds. This gene was characterized by Geisen et al. (2006) and belongsto a gene cluster of the OTA biosynthethic pathway in Penicillium(otanpsPN, accession number AY557343). In addition, it has beendemonstrated that the presence of the non-ribosomal peptidesynthethase is necessary for OTA biosynthesis in both Penicillium andAspergillus species (Geisen et al., 2006, O'Callaghan et al., 2006). Primersand probe designed in the present work match with all the obtainedsequences from otanpsPN gene for all OTA-producing species. Inaddition, the primer pairwas appropriate for qPCR, since primer-dimerswere not detected. The small divergence in the Tmvalue (84.0–84.9 °C)could be due to small differences in the sequence of the target otanpsPNgene of the different species and genera analyzed given that the DNAmelting curves are a function of the GC/AT radio, length, and sequences(Ririe et al., 1997). This was confirmed by sequence data which showedup to 16 variable nucleotide sites when comparing sequences of theotanpsPN gene from the different species and genera used in the presentstudy.
The specificity of the designed primers pair F/R-npstr and probeNPSprobe was confirmed in this study since the results obtained fromboth qPCR protocols were closely related to the OTA detection byMECEand HPLC-MS. Furthermore, high positive correlation between the Ctvalues obtained by developed qPCRmethods andOTA production of thetested strains was observed (0.84 with SYBR Green and 0.85 withTaqMan assay). All OTA-producing strains detected by above methodsfrom the three genera tested (Penicillium, Aspergillus and Emericella)gave a positive result in each optimized assay. No amplification by qPCRwas observed for any of the non-OTA producers. In addition, severalspecies from Penicillium (P. carneum, P. melanoconidium andP. aurantiogriseum) and Aspergillus (A. tamarii) genus which previouslyhave never been described as ochratoxin A producers showedproduction of this mycotoxin by MECE and HPLC-MS. In all of thesespecies a partial sequences of otanpsPN gene matching with thepublished sequences of otanpsPN gene of P. nordicum (GeneBankaccession no. AY557343) was obtained. This confirms that they areproducing species. In addition, to check the identification with thatreported by their respective Culture Collection of precedence, theformer specieswere identifiedby amplificationpartial sequencingof ITSregion and β-tubulin gene (GenBank accession nos. JN097809;JN097810; JN097811; JN097812; JN097813; JN097814, JN217227,JN217233). In all cases the identification agrees with that reported byculture collection. From these results, the two optimized methods canbe considered specific and provided a good discrimination betweenOTA-producing and non-producing strains across species and genera.
Table 4Ct values of pure DNA of 3 OTA producing species in deionized water and in DNA tracted from different amounts of food matrices, respectively, using SYBR Green and Taq an protocols.
qPCR method OTA producing reference strains
P. aurantiogriseum CECT 2264 P. nordicum CBS 110769 A. ochraceus CBS 89.68
Pure DNA+deionized water
Pure DNA+total DNA extracted fromdifferent amounts of cooked ham (g)
Pure DNA+deionized water
Pure DNA+total DNA extracted fromdifferent amounts of ripened cheese (g)
Pure DNA+dei ized water Pure DNA+total DNA extracted fromdifferent amounts of grape (g)
aData represent the mean threshold cycle (Ct)±standard deviation (SD) of the 5 dependent DNA extractions. No significant differences (pb0.05) were found between Ct v ued of pure DNA and pure DNA spiked with DNA from food.No Template Control (NTC) Ct value was established in 34–35 for SYBR Green and aqMan assay.
Table 5Quantification of conidia from OTA producing molds in artificially inoculated food SYBR Green and TaqMan qPCR.
Level of Inoculatedconidia (conidia·g−1)
qPCR method Food products
aCooked products bRipened products Fruits
Mortadella Cooked turkeybreast
Cooked ham Dry cured ham Dry fermentedsausage “Salchichón”
a Data extrapolated from standard curve for cooked ham inoculated with P. aura tiogriseum CECT 2264.b Data extrapolated from standard curve for ripened cheese inoculated with P.n dicum CBS 110769.c Data extrapolated from standard curve for grape inoculated with A.ochraceus C S 589.68.d Data represent the mean amount of conidia quantified±standard deviation of he 3 independent assays, each consisting of triplicate samples.
233A.Rodríguez
etal./
InternationalJournalof
FoodMicrobiology
149(2011)
226–235
ex
1231
inT
by
3
3
norBt
M
5
on
al
c
G
((((
234 A. Rodríguez et al. / International Journal of Food Microbiology 149 (2011) 226–235
Furthermore, it was demonstrated that the presence of 10 ng of non-targeted DNA in the qPCR did not interfere with the detection of smallamounts (0.1 pg) of target DNA.
With regard to sensitivity of the optimized qPCR methods, thelimits of detection, determined as the smallest amount of fungal DNAdetected, were 0.01 pg for SYBR Green and TaqMan assays when pureDNA was used. The latter protocol was 10 times more sensitive thanother qPCR methods described in the literature for detection oftoxigenic molds (Mayer et al., 2003; Morello et al., 2007). In addition,the sensitivity was also evaluated in different food matrices. The limitof detection in all inoculated foods ranged from 10 to 1 conidia/g forboth optimized qPCRmethods, andwas lower than previously reported(Selma et al., 2008).
Although no guidelines have been established for standard curvesused in qPCR assays that measure fungi, Fredlund et al. (2008) andSuanthie et al. (2009) suggested the use of criteria established forGenetically Modified Organism analysis of foods where the slope of thestandard curve should range between−3.1 and−3.6 and the R2 valuebe≥0.98. These guidelines should also be valid for the analysis of fungalDNA in foods. In the present work, the two optimized methods frompure DNA had R2 values around 0.99 and showed appropriate slope-values within the optimum range. These results indicated that bothqPCR methods could be applied for quantifying OTA producers directlyin foods.
This fact was confirmed with standard curves obtained from allinoculated foods. All standard curves showed good linearity (R2 N0.98)and their slopes were within the recommended range. The efficiency ofthe standard curve for A.ochraceus CBS 589.68 and A. westerdijkiae CECT2948 in grapes using both qPCR assays were lower than observed inother inoculated food matrices, probably due to the presence of PCR-inhibiting substances such as tannins, polysaccharides and pigments(Mulé et al., 2006; Selma et al., 2008). Furthermore, no amplificationfrom non-producing strain A. westerdijkiae 3.38 in inoculated foodsdemonstrates the robustness of the qPCR assays.
To test the ability of the qPCR methods to quantify OTA-producingmolds in foods, two known amounts of fungal conidia were added todifferent foods and quantified by the qPCR methods. Standard de-viations observed in the present study where conidia were addeddirectly to the food matrix were comparable to previous studies whereconidia were added to food extracts (Selma et al., 2008). Similar valuesof standard deviations have been reported for other qPCR protocolsdeveloped to detect mycotoxin-producing molds when DNA wasextracted fromartificially inoculated or naturally contaminated samples(Atoui et al., 2007;González-Salgado et al., 2009;Nicolaisenet al., 2009).
Several compounds in fruits and ripened products may influencequantification (Monnetet al., 2006,Muléet al., 2006) by interferingwithPCR amplification (Demeke and Jenkins 2010; Fredlund et al., 2008).However, no differences in the quantification of OTA-producing moldsbetween the food matrices have been found in the present study.
Regarding accuracy of the developed methods when applied tofoods, themean values of conidia quantifiedwith TaqManmethodologywere closer to the inoculated amount than those with the SYBR Greenassay. Nevertheless, the sensitivity was higher with SYBR Green thanwith the TaqMan methodology. In addition, the SYBR Green method issimpler than TaqMan method, since no probe is required and this mayreduce costs in routine food analyses (Cutrín et al., 2009).
Both SYBR Green and TaqMan qPCR procedures developed in thepresent study to quantify OTA-producing molds in food, could beperformed within a relatively short time (5–6 h for DNA extraction and2–3 h for qPCR). This time is considerably lower than that needed toquantify OTA producing mold strains by conventional culturing tech-niques (3–5 days).
In conclusion, the qPCR based procedures developed in this studyshow promise for use for specific detection and quantification of OTA-producing molds directly in food samples and to prevent OTA fromentering the food chain.
Acknowledgements
This work has been funded by project AGL2007-64639 of theSpanish Comision Interministerial de Ciencia y Tecnología, CarnisenusaCSD2007-00016, Consolider Ingenio2010andGRU08100andGRU09158of the Junta de Extremadura and FEDER. The work of AnnemarieF. Justesen was supported by the Danish Food Industry Agency, project3412-07-01873. Alicia Rodríguez would like to thank the SpanishComision Interministerial de Ciencia y Tecnología for the pre-doctoralgrant (BES-2008-008021) and for the pre-doctoral grant for short stagein Aarhus University (Denmark).
References
Abarca, M.L., Accensi, F., Bragulat, M.R., Castella, G., Cabanes, F.J., 2003. Aspergilluscarbonarius as the main source of ochratoxin A contamination in dried vine fruitsfrom the Spanish Market. Journal of Food Protection 66, 504–506.
Abbas, A., Valez, H., Dobson, A.D.W., 2009. Analysis of the effect of nutritional factors onOTA and OTB biosynthesis and polyketide synthase gene expression in Aspergillusochraceus. International Journal of Food Microbiology 135, 22–27.
Atoui, A., Mathieu, F., Lebrihi, A., 2007. Targeting a polyketide synthase gene forAspergillus carbonarius quantification and ochratoxin A assessment in grapes usingreal-time PCR. International Journal of Food Microbiology 115, 313–318.
Ayoub, F., Reverberi, M., Ricelli, A., D'Onghia, A.M., Yaseen, T., 2010. Early detection ofAspergillus carbonarius and A. niger on table grapes: a tool for quality improvement.Food Additives and Contaminant 27, 1285–1293.
Bellí, N., Pardo, E., Marín, S., Farré, G., Ramos, A.J., Sanchís, V., 2004. Occurrence ofochratoxin A and toxigenic potential of fungal isolates from Spanish grapes. Journalof the Science of Food and Agriculture 84, 541–546.
Bogs, C., Battilani, P., Geisen, R., 2006. Development of a molecular detection anddifferentiation system for ochratoxin A producing Penicillium species and itsapplication to analyse the occurrence of Penicillium nordicum in cured meats.International Journal of Food Microbiology 107, 39–47.
Cutrín, J.M., Olveira, J.G., Bandín, I., Dopazo, C.P., 2009. Validation of real-time RT-PCRapplied to cell culture for diagnosis of any known genotype of viral haemorrhagicsepticaemia virus. Journal of Virological Methods 162, 155–162.
Dachoupakan, C., Ratomahenina, R., Martinez, V., Guiraud, J.P., Baccou, J.C., Schorr-Galindo, S., 2009. Study of the phenotypic and genotypic biodiversity of potentiallyochratoxigenic black aspergilli isolated from grapes. International Journal of FoodMicrobiology 132, 14–23.
Demeke, T., Jenkins, R., 2010. Influence of DNA extraction methods, PCR inhibitors andquantification methods on real-time PCR assay of biotechnology-derived traits.Analytical and Bioanalytical Chemistry 396, 1977–1990.
Fredlund, E., Gidlund, A., Olsen, M., Börjesson, T., Spliid, N.H.H., Simonsson, M., 2008.Method evaluation of Fusarium DNA extraction from mycelia and wheat for down-stream real-time PCR quantification and correlation to mycotoxin levels. Journal ofMicrobiological Methods 73, 33–40.
Gallo, A., Perrone, G., Solfrizzo, M., Epifani, F., Abbas, A., Dobson, A.D.W., Mulé, G., 2009.Characterization of a pks gene which is expressed during ochratoxin A productionby Aspergillus carbonarius. International Journal of Food Microbiology 129, 8–15.
Geisen, R., Schmidt-Heydt, M., Karolewiez, A., 2006. A gene cluster of the ochratoxin Abiosynthetic genes in Penicillium. Mycotoxin Research 22, 134–141.
Gil-Serna, J., Gónzalez-Salgado, A., Gónzalez-Jaén, M.T., Vázquez, C., Patiño, B., 2009.ITS-based detection and quantification of Aspergillus ochraceus and Aspergilluswesterdijkiae in grapes and green coffee beans by real-time quantitative PCR.International Journal of Food Microbiology 131, 162–167.
Gil-Serna, J., Vázquez, C., Sardiñas, N., González-Jaén, M.T., Patiño, B., 2011. Revision ofochratoxin A production capacity by the main species of Aspergillus sectionCircumdati. Aspergillus steynii revealed as the main risk of OTA contamination.Food Control 22, 343–345.
Glass, N.L., Donaldson, G.C., 1995. Development of primer sets designed for use with thePCR to amplify conserved genes from filamentous Ascomycetes. Applied andEnvironmental Microbiology 61, 1323–1330.
González-Salgado, A., Patiño, B., Gil-Serna, J., Vázquez, C., González-Jaén, M.T., 2009.Specific detection of Aspergillus carbonarius by SYBR® Green and TaqMan®quantitative PCR assays based on themulticopy ITS2 region of the rRNA gene. FEMSMicrobiology Letters 295, 57–66.
Harris, J.P., Mantle, P.G., 2001. Biosynthesis of ochratoxins by Aspergillus ochraceus.Phytochemistry 58, 709–716.
Huff, W.E., Hamilton, P.B., 1979. Mycotoxins—their biosynthesis in fungi: ochratoxins—metabolites of combined pathways. Journal of Food Protection 42, 815–820.
Karolewiez, A., Geisen, R., 2005. Cloning a part of the ochratoxin A biosynthetic genecluster of Penicillium nordicum and characterization of the ochratoxin polyketidesynthase gene. Systematic and Applied Microbiology 28, 588–595.
Kubista, M., Andrade, J.M., Bengtsson, M., Forootan, A., Jonák, J., Lind, K., Sindelka, R.,Sjöback, R., Sjögreen, B., Strömbom, L., Stahlberg, A., Zoric, N., 2006. The real-timepolymerase chain reaction. Molecular Aspects of Medicine 27, 95–125.
Kure, C.F., Skaar, I., Brendehaug, J., 2004. Mould contamination in production of semi-hard cheese. International Journal of Food Microbiology 93, 41–49.
López-Mendoza, M.C., Crespo-Sempere, A., Martínez-Culebras, P.V., 2009. Identificationof Aspergillus tubingensis strains responsible for OTA contamination in grapes and
235A. Rodríguez et al. / International Journal of Food Microbiology 149 (2011) 226–235
wine based on the acyl transferase domain of a polyketide synthase gene.International Journal of Food Science and Technology 44, 2147–2152.
Lund, F., Frisvad, J.C., 2003. Penicillium verrucosum in wheat and barley indicatespresence of ochratoxin A. Journal of Applied Microbiology 95, 1117–1123.
Martín, A., Jurado, M., Rodríguez, M., Núñez, F., Córdoba, J.J., 2004. Characterization ofmolds from dry-cured meat products and their metabolites by micellar electroki-netic capillary electrophoresis and random amplified polymorphic DNA PCR.Journal of Food Protection 67, 2234–2239.
Mayer, Z., Bagnara, A., Färber, P., Geisen, R., 2003. Quantification of the copy number ofnor-1, a gene of the aflatoxin biosynthetic pathway by real-time PCR, and itscorrelation to the cfu of Aspergillus flavus in foods. International Journal of FoodMicrobiology 82, 143–151.
MikuŠovà, P., Ritieni, A., Santini, A., Juhasov , G., Šrob rov , A., 2010. Contamination bymoulds of grape berries in Slovakia. Food Additives and Contaminant 27, 738–747.
Monnet, C., Correia, K., Sarthou, A.S., Irlinger, F., 2006. Quantitative detection ofCorynebacterium casei in cheese by real-time PCR. Applied and EnvironmentalMicrobiology 72, 6972–6979.
Morello, L.G., Sartori, D., de Oliveira Martínez, A.L., Vieira, M.L.C., Taniwaki, M.H.,Fungaro, M.H.P., 2007. Detection and quantification of Aspergillus westerdijkiae incoffee beans based on selective amplification of β-tubulin gene by using real-timePCR. International Journal of Food Microbiology 119, 270–276.
Mulé, G., Susca, A., Logrieco, A., Stea, G., Visconti, A., 2006. Development of aquantitative real-time PCR assay for the detection of Aspergillus carbonarius ingrapes. International Journal of Food Microbiology 111, S28–S34.
Nicolaisen, M., Suproniené, S., Nielsen, L.K., Lazzaro, I., Spliid, N.H., Justesen, A.F., 2009.Real-time PCR for quantification of eleven individual Fusarium species in cereals.Journal of Microbiological Methods 76, 234–240.
Núñez, F., Rodríguez, M.M., Bermúdez, E., Córdoba, J.J., Asensio, M.A., 1996. Compositionand toxigenic potential of the mould population on dry-cured Iberian ham.International Journal of Food Microbiology 32, 185–197.
O'Callaghan, J., Stapleton, P.C., Dobson, A.D.W., 2006. Ochratoxin A biosynthetic genesin Aspergillus ochraceus are differentially regulated by pH and nutritional stimuli.Fungal Genetics and Biology 43, 213–221.
Ririe, K.M., Rasmussen, R.P., Wittwer, C.T., 1997. Product differentiation by analysis ofDNA melting curves during the polymerase chain reactions. Analytical Biochem-istry 245, 154–160.
Sánchez, B., Rodríguez, M., Casado, E.M., Martín, A., Córdoba, J.J., 2008. Development ofan efficient fungal DNA extraction method to be used in random amplifiedpolymorphic DNA–PCR analysis to differentiate cyclopiazonic acid mold producers.Journal of Food Protection 71, 2497–2503.
Sartori, D., Furlaneto, M.C., Martins, M.K., Ferreira de Paula, M.R., Pizzirani-Kleiner, A.A.,Taniwaki, M.H., Pelegrini Fungaro, M.H., 2006. PCR method for the detection ofpotential ochratoxin-producing Aspergillus species in coffee beans. Research inMicrobiology 157, 350–354.
Schmidt-Heydt, M., Baxter, E., Geisen, R., Magan, N., 2007. Physiological relationshipbetween food preservatives, environmental factors, ochratoxin and otapksPV geneexpression by Penicillium verrucosum. International Journal of Food Microbiology119, 277–283.
Selma, M.V., Martínez-Culebras, P.V., Aznar, R., 2008. Real-time PCR based proceduresfor detection and quantification of Aspergillus carbonarius in wine grapes.International Journal of Food Microbiology 122, 126–134.
Sosa, M.J., Córdoba, J.J., Díaz, C., Rodríguez, M., Bérmudez, E., Asensio, M.A., Núñez, F.,2002. Production of cyclopiazonic acid by Penicillium commune isolated from dry-cured ham on a meat extract-based substrate. Journal of Food Protection 65,988–992.
Suanthie, Y., Cousin, M.A., Woloshuk, C.P., 2009. Multiplex real-time PCR for detectionand quantification ofmycotoxigenic Aspergillus, Penicillium and Fusarium. Journal ofStored Products Research 45, 139–145.
Sweeney, M.J., Dobson, A.D.W., 1998. Mycotoxin production by Aspergillus, Fusariumand Penicillium species. International Journal of Food Microbiology 43, 141–158.
Thirumala-Devi, K., Mayo, M.A., Reddy, G., Emmanuel, K.E., Larondelle, Y., Reddy, D.V.R.,2001. Occurrence of ochratoxin A in black pepper, coriander, ginger and turmeric inIndia. Food Additives and Contaminants 18, 830–835.
White, T.J., Burns, T., Lee, S., Taylor, J.W., 1990. Amplification and direct sequencing offungal ribosomal RNA genes for phylogenetics. In: Innis, M.A., Gelgard, D.H.,Sninsky, J.J., White, T.J. (Eds.), PCR Protocols: A Guide to Methods and Applications.Academic Press, New York, pp. 315–322.
Wilson, L.G., 1997. Inhibition and facilitation of nucleic acid amplification. Applied andEnvironmental Microbiology 63, 3741–3751.
