International Journal of Food Microbiology 165 (2013) 111–120

Contents lists available at SciVerse ScienceDirect

International Journal of Food Microbiology

j ourna l homepage: www.e lsev ie r .com/ locate / i j foodmicro

Genetic structure and natural variation associated with host of origin inPenicillium expansum strains causing blue mould

S.M. Sanzani a,⁎, C. Montemurro a, V. Di Rienzo a, M. Solfrizzo b, A. Ippolito a

a Department of Soil, Plant and Food Sciences, University of Bari “Aldo Moro”, Via G. Amendola 165/A, 70126 Bari, Italyb Institute of Sciences of Food Production, National Research Council, Via G. Amendola 122/O, 70126 Bari, Italy

⁎ Corresponding author. Tel.: +39 0805443055; fax:E-mail address: simonamarianna.sanzani@uniba.it (

0168-1605/$ – see front matter © 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.ijfoodmicro.2013.04.024

a b s t r a c t

a r t i c l e i n f o

Article history:Received 20 December 2012Received in revised form 10 April 2013Accepted 18 April 2013Available online 1 May 2013

Keywords:Penicillium expansumPatulinHigh Resolution MeltingHost specificityPathogenicityVirulence

Blue mould, caused by Penicillium expansum, is one of the most economically damaging postharvest diseasesof pome fruits, although it may affect a wider host range, including sweet cherries and table grapes. Severalreports on the role of mycotoxins in plant pathogenesis have been published, but few focussed on theinfluence of mycotoxins on the variation in host preference amongst producing fungi. In the present studythe influence of the host on P. expansum pathogenicity/virulence was investigated, focussing mainly on therelationship with patulin production. Three P. expansum strain groups, originating from apples, sweet cherries,and table grapes (7 strains per host) were grown on their hosts of isolation and on artificial media derivedfrom them. Strains within each P. expansum group proved to be more aggressive and produced more patulinthan the other two groups under evaluation when grown on the host from which they originated. Table grapestrains were the most aggressive (81% disease incidence) and strongest patulin producers (up to 554 μg/g).The difference in aggressiveness amongst strainswas appreciable only in the presence of a living host, suggestingthat the complex pathogen–host interaction significantly influenced the ability of P. expansum to cause thedisease. Incidence/severity of the disease and patulin production proved to be positively correlated, supportingthe role of patulin as virulence/pathogenicity factor. The existence of genetic variation amongst isolates wasconfirmed by the High Resolution Melting method that was set up herein, which permitted discrimination ofP. expansum from other species (P. chrysogenum and P. crustosum) and, within the same species, amongst thehost of origin. Host effect on toxin production appeared to be exerted at a transcriptional level.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Bluemould is considered one of themost important postharvest dis-eases of pome fruit worldwide (Pianzzola et al., 2004), although it cancause serious decay amongst a wider range of harvested commoditiessuch as stone fruit, soft fruit and berry fruit (Neri et al., 2010). The dis-ease may be caused by various Penicillium species, of which Penicilliumexpansum is the most aggressive and commonly encountered. Amongstother species, P. solitum, P. crustosum, and P. chrysogenum have beenreported (Pitt et al., 1991; Sanderson and Spotts, 1995). Blue mould,being a soft rot, severely affects the quality properties of the infectedfruit making it unmarketable, with consequent economic losses forretailers. Furthermore, species in genus Penicillium are well knownproducers of pharmaceutically active compounds but also dangerousmycotoxins (Frisvad and Samson, 2004). For example, P. expansumhas been shown to produce several toxic compounds including patulin,whichhasmutagenic, immunotoxic, and neurotoxic properties (Castoriaet al., 2012), so that its content in apple-derived products has been reg-ulated by the European Commission (2006). Since patulin is extremely

+39 0805442911.S.M. Sanzani).

rights reserved.

stable at acidic pH and resistant to pasteurization, it is critical for thefruit processing industry to detect and minimize P. expansum rots infruits destined to juice production. However, the effectiveness of thecontrol is often reduced by several pathogen characteristics such as:i) the development of resistance to the few fungicides permitted in thepostharvest phase; ii) the penetration through wounds (producedduring picking and handling operations), natural openings (stem end,open calyx tube, lenticels, etc.) or infection sites of other primarypathogens; iii) the ability to grow both at refrigeration temperaturesused for storage (−1, 0 °C) and in warmer environments at retailand consumer sites, particularly on over-mature/long-stored fruits(Mari et al., 2009). Most studies on patulin are focussed on applesand derived products, since apple fruits are reported as the mostsusceptible and the only ones for which regulatory limits havebeen imposed (Sant'Ana et al., 2008). However, patulin productionat considerable levels has been reported even on other fruit hosts(Larsen et al., 1998; Neri et al., 2010; Piemontese et al., 2005;Reddy et al., 2010). Therefore, it is conceivable that in the future,patulin contamination of fresh fruits other than apples and theirderived products might become a bigger issue for the food industry.

The role of mycotoxins in plant pathogenesis is still not completelyunderstood. In general, secondary metabolites produced by fungi are

Table 1Penicillium spp. strains, host of isolation, species, patulin production (on PDA) and SNP-cluster. In bold the selected P. expansum strains and the accession numbers of partialβ-tubulin sequences deposited in GenBank are reported.

Isolate name Host Identification Patulin production (μg/cm2) SNP-cluster Accession n.

Pex1 Apple Penicillium crustosum 0 6Pex2 Cherry Penicillium expansum 4.23 1Pex3 Cherry Penicillium expansum 4.62 1Pex4 Apple Penicillium expansum 11.37 2 KC342827Pex5 Apple Penicillium expansum 2.45 2 KC342828Pex6 Apple Penicillium expansum 13.11 2 KC342829Pex7 Apple Penicillium expansum 6.26 2 KC342830Pex8 Apple Penicillium expansum 8.30 1Pex9 Cherry Penicillium expansum 12.47 1Pex10 Pear Penicillium crustosum 0 4Pex11 Cherry Penicillium chrysogenum 0 3Pex12 Almond Penicillium crustosum 0 5Pex13 Cherry Penicillium expansum 5.09 1Pex14 Cherry Penicillium expansum 7.19 1Pex15 Cherry Penicillium expansum 6.44 1 KC342831Pex16 Cherry Penicillium expansum 6.40 1Pex17 Cherry Penicillium expansum 4.53 1Pex18 Cherry Penicillium expansum 7.53 1 KC342832Pex19 Cherry Penicillium expansum 12.96 1 KC342833Pex20 Cherry Penicillium expansum 0.60 1Pex21 Almond Penicillium chrysogenum 0 3Pex22 Cherry Penicillium expansum 5.76 1Pex23 Cherry Penicillium expansum 6.98 1Pex24 Cherry Penicillium expansum 5.18 1Pex25 Cherry Penicillium expansum 6.41 1Pex26 Cherry Penicillium expansum 6.63 1Pex27 Cherry Penicillium chrysogenum 0 3Pex28 Apricot Penicillium chrysogenum 0 3Pex29 Apricot Penicillium chrysogenum 0 3Pex30 Almond Penicillium crustosum 0 1Pex31 Cherry Penicillium expansum 10.18 1 KC342834Pex32 Cherry Penicillium expansum 4.38 1Pex33 Cherry Penicillium expansum 8.34 1 KC342835Pex34 Cherry Penicillium expansum 6.06 1Pex35 Cherry Penicillium expansum 6.07 1Pex36 Cherry Penicillium expansum 7.60 1 KC342836Pex37 Cherry Penicillium expansum 6.06 1Pex38 Cherry Penicillium expansum 5.61 1Pex39 Cherry Penicillium expansum 6.34 1Pex40 Cherry Penicillium expansum 7.39 1Pex41 Cherry Penicillium expansum 11.30 1 KC342837Pex42 Cherry Penicillium expansum 7.82 1Pex43 Cherry Penicillium expansum 8.17 1Pex44 Cherry Penicillium crustosum 0 4Pex45 Apple Penicillium expansum 6.30 1Pex46 Cherry Penicillium expansum 3.32 1Pex47 Apple Penicillium expansum 7.71 1 KC342838Pex48 Grape Penicillium expansum 10.25 1Pex49 Grape Penicillium expansum 9.11 1Pex50 Grape Penicillium expansum 9.79 1Pex51 Grape Penicillium expansum 0.00 1Pex52 Grape Penicillium expansum 7.26 1Pex53 Grape Penicillium expansum 6.87 1Pex54 Grape Penicillium expansum 21.22 1Pex55 Grape Penicillium expansum 13.47 1Pex56 Grape Penicillium expansum 8.66 1Pex57 Grape Penicillium expansum 6.34 1Pex58 Grape Penicillium expansum 6.82 1Pex59 Raspberry Penicillium crustosum 0 5Pex60 Grape Penicillium expansum 21.56 1 KC342839Pex61 Grape Penicillium expansum 28.48 1 KC342840Pex62 Grape Penicillium expansum 26,17 1Pex63 Grape Penicillium expansum 27,67 1 KC342841Pex64 Grape Penicillium expansum 12.20 1Pex65 Grape Penicillium expansum 15.58 1Pex66 Grape Penicillium expansum 19.96 1Pex67 Grape Penicillium expansum 24.94 1 KC342842Pex68 Grape Penicillium expansum 10.10 1Pex69 Grape Penicillium expansum 12.93 1Pex70 Grape Penicillium expansum 15.91 1Pex71 Grape Penicillium expansum 16.84 1Pex72 Grape Penicillium expansum 18.45 1 KC342843Pex73 Grape Penicillium expansum 0.76 1Pex74 Grape Penicillium expansum 0.75 1

112 S.M. Sanzani et al. / International Journal of Food Microbiology 165 (2013) 111–120

Table 1 (continued)

Isolate name Host Identification Patulin production (μg/cm2) SNP-cluster Accession n.

Pex75 Grape Penicillium expansum 17.82 1Pex76 Grape Penicillium expansum 17.48 1Pex77 Grape Penicillium expansum 27.42 1 KC342844Pex78 Grape Penicillium expansum 18.59 1Pex79 Grape Penicillium expansum 18.39 1 KC342845Pex80 Grape Penicillium expansum 13.34 1Pex81 Grape Penicillium expansum 8.19 1Pex82 Grape Penicillium expansum 38.40 1Pex83 Raspberry Penicillium crustosum 0 5Pex84 Apple Penicillium expansum 11.79 2 KC342846Pex85 Apple Penicillium expansum 8.55 1Pex86 Apple Penicillium expansum 10.22 1 KC342847Pex87 Pear Penicillium crustosum 0 4

113S.M. Sanzani et al. / International Journal of Food Microbiology 165 (2013) 111–120

not required for their growth, but presumably can confer some selectiveadvantage, for example protecting mycotoxin-producing fungi againstother organisms sharing the same trophic niche or exerting a toxiceffect on the host (Reverberi et al., 2010). Furthermore, it has beenreported that the plant may influence toxin production by means ofnatural constituents or signal molecules produced as a response topathogen attack (Howlett, 2006). For instance, in some Aspergillusspp., such as A. flavus, A. parasiticus and A. nidulans, the acetyl CoAmolecules necessary for the synthesis of sterigmatocystin and aflatoxinsare derived from fatty acids present in the colonized maize kernels(Maggio-Hall et al., 2005).

It has been reported that patulin producing P. expansum strainshave a higher frequency of isolation, level of virulence, growth rateand competitive ability over other non producing Penicillium spp.such as P. solitum (Sanderson and Spotts, 1995). Moreover, the impor-tance of patulin for P. expansum pathogenicity and virulence has beensupported by recent experimental evidence (Sanzani et al., 2012).However, to the best of our knowledge, the role played by the hostin influencing the production of patulin as pathogenesis-related fac-tor and consequently P. expansum aggressiveness, has not been clear-ly established.

In the present investigation, a selection of P. expansum strainscausing blue mould, isolated from different hosts (apple, sweet cherryand table grape), were evaluated for their growth and toxin produc-tion ability in vivo and in vitro on the hosts of isolation and theirderived artificial media, respectively. Furthermore, the influence ofthe host composition/signal molecules on the patulin biosyntheticpathway was analysed from a transcriptomic and genetic aspect.This latter task was pursued by means of High Resolution Melting(HRM), which is the analysis of the melt curve of a DNA fragmentfollowing amplification by Real-time PCR. Indeed, HRM experimentsgenerate melt curve profiles that are both specific and sensitiveenough to detect even single base variations (Garritano et al., 2009),paving the way for small- and large-scale genotyping of populationsof any organism. The evolution of HRM from conventional meltingcurve analysis is based on the development of a novel type of dye,having a higher dsDNA saturating activity than SYBR Green I, calledEvaGreen®, and of a new generation of thermal cyclers, finely tuningtemperature increments/decrements up to 0.1 °C (Gori et al., 2012).

The final aim of this work was to gain insight into the existence of“specialization” amongst P. expansum isolates for a particular hostrelated to the host's influence on patulin production and Penicilliumpathogenicity/virulence. This, in turn, will further confirm patulin asa pathogenesis-related factor.

2. Materials and methods

2.1. Penicillium isolates tested

Over time a freely available collection of 87 Penicillium spp. isolateshas been built up at the Department of Soil, Plant and Food Sciences of

the University of Bari “Aldo Moro”. Fruits (almonds, apples, apricots,sweet cherries, table grapes, pears, and raspberries) with visual signsof Penicillium spp. contamination were carefully placed in plastic bagsand incubated in the dark at 24 °C for 4 days (Pitt and Hocking,2009). Conidia from each lesionwere collected and suspended in steriledistilledwater containing 0.01% (v/v) Tween80. Suspensionswere thenspread on semi-selective potato dextrose agar (PDA), prepared fromfresh potatoes as reported by Samson and Pitt (2000) and amendedwith ampicillin (250 mg/l, Sigma-Aldrich, St. Louis, USA). Dishes werethen incubated at 24 °C for 2 days and, for eachof them, themost prom-inent amongst the putative Penicillium spp. emerging colonies wastransferred to a fresh PDA plate. All cultures were incubated for7 days at 24 °C and purified as required into monoconidial isolates(Pitt and Hocking, 2009). The collected fungal isolates were recordedas reported in Table 1, classified according to macro- and micromor-phology features as described by Frisvad and Samson (2004), and eval-uated for their patulin production ability on PDA as reported by Sanzaniet al. (2009a). Results were expressed as μg of patulin per cm2 of dishsurface.

2.2. Molecular characterization of Penicillium strains

Twenty-one isolates (Table 1, in bold), 7 for each of the most rep-resented hosts of isolation (apples, sweet cherries, and table grapes)in the Penicillium spp. collection, were selected for their patulin pro-duction potential on PDA and molecularly characterized.

2.2.1. DNA extractionDNA was isolated from approximately 100 mg of mycelium col-

lected from PDA plates according to Schena et al. (2002) and RNaseAtreated as reported by the manufacturer (Sigma-Aldrich). Its concen-tration, purity and integrity were determined by a spectrophotometerand by electrophoresis on a 0.7% agarose gel in TAE buffer accordingto standard procedures (Sambrook and Russell, 2001).

2.2.2. Species-specific primersDNA from Penicillium isolateswas employed in PCR reactions carried

out according to Marek et al. (2003) by using P. expansum-specificprimers (PEF 5′-ATCGGCTGCGGATTGAAAG-3′ and PER 5′-AGTCACGGGTTTGGAGGGA-3′), that amplify a 404 bp fragment targeting thepolygalacturonase gene. Electrophoresis on a 1.2% agarose gel wasused to detect the presence of the PCR product from each isolate.

2.2.3. Gene-sequencingA fragment of about 450 bp of the β-tubulin gene was amplified

using primers Bt2a (5′-GGTAACCAAATCGGTGCTGCTTTC-3′) and Bt2b(5′-ACCCTCAGTGTAGTGACCCT TGGC-3′) according to Glass andDonaldson (1995). The band of the expected size was excisedfrom the gel and purified with GenElute Gel Extraction Kit (Sigma-Aldrich). Sequencing reactions were performed by Primm srl (Milan,Italy). The obtained nucleotide sequences were submitted to the online

114 S.M. Sanzani et al. / International Journal of Food Microbiology 165 (2013) 111–120

BLAST search engine of the National Centre for BiotechnologyInformation (NCBI), using the N or X algorithms, to search for similarityof the query according to the degree of sequence match. The sequenceswere deposited inGenBank (accession numbers are reported in Table 1)and aligned using the freely available softwareMULTALIN (http://npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_multalin.html). The presence of single nucleotide polymorphisms (SNPs) wasevaluated.

2.3. High Resolution Melting (HRM) analysis

On the consensus β-tubulin sequence fragment obtained bythe alignment and including the SNPs, the primer pair HRMF 5′-ACGTCTACTTCAACCATGTG-3′ and HRMR 5′-CACCGCTGGCCTAGATTATC-3′ was designed using the software Oligo Explorer 1.1.2(http://www.genelink.com/).

The reaction efficiency and correlation coefficient were deter-mined by amplifying different amounts of DNA template (102 to10−2 ng) in Real-time PCR reactions. Standard curves and linearequations were built up using the instrument associate software byplotting Cq values (y-axis) against logs of DNA (x-axis). Reactionspecificity was confirmed by melt curve analysis and by electrophore-sis of the amplification product on 1.5% agarose gel.

Reactions were carried out in a CFX96 cycler-Real-time PCRDetection System (Bio-Rad Laboratories Inc., Hercules, CA, USA), in96 wells PCR plates. Ten μl reaction mixture contained 0.5 mM ofeach primer, 1× SsoFast EvaGreenH Supermix (Bio-Rad), and 15 ngof genomic DNA as template, unless otherwise specified. HRM analy-ses included all the isolates present in the Penicillium spp. collection(Table 1). Cycling conditions were: one cycle at 95 °C for 5 min,followed by 40 cycles at 94 °C for 20 s, 55 °C for 20 s and 72 °C for20 s. The amount of fluorescence for each sample was measured atthe end of each cycle and analyzed via CFX-Manager Software v1.6(Bio-Rad). Melting curves of PCR amplicons were obtainedwith tem-peratures ranging from 65 °C to 95 °C. Acquisition was performedevery 0.5 °C (0.2 °C for HRM reactions) increase in temperature,with a 10 s step.

Data were analyzed by High Resolution Melting analysis software(Bio-Rad), which automatically clustered the samples according totheir melting profiles. This clustering profile was then confirmed bysequencing, as reported above, a representative amplification productfor each cluster. The obtained sequences were analysed by the freelyavailable softwares ClustalX2 and TreeView X and a phylogenetic treewas carried out.

2.4. Growth and patulin production on different culture media and hosts

2.4.1. Inoculum preparationSeven P. expansum PDA cultures per host were washed with 6 ml

of sterile distilled water containing 0.05% (v/v) Tween 80. Eachresulting spore suspension was filtered through two layers of sterilegauze and spore counts were made by a Thoma counting chamber(HGB Henneberg-Sander GmbH, Lutzellinden, Germany). A dilutedconidial suspension with a final concentration of 104 conidia/ml foreach strain was used for all in vitro and in vivo trials.

2.4.2. In vitro assayTo test the effect of the host composition on the growth and patulin

production of the strains under evaluation, they were grown on differ-ent artificial media made with their hosts of isolation. These mediawere prepared by replacing potatoes in PDA recipewith 200 g of apples(cv. Golden Delicious, ADA), sweet cherries (cv. Ferrovia, CDA) andtable grapes (cv. Italia, GDA). Dishes were centrally inoculated with10 μl of conidial suspension and incubated at 24 °C in the dark for7 days. For each strain (7 per host of isolation, 21 in total) andmedium,tests were performed in triplicate (three agar dishes). At the end of

incubation colony diameter (mm) and patulin production (μg/cm2)were recorded. Patulin was extracted from P. expansum mycelium asreported above.

2.4.3. In vivo assayFreshly harvested apples (cv. Golden Delicious), sweet cherries

(cv. Ferrovia) and table grapes (cv. Italia), free of defects or injuriesand uniform in size, colour and ripeness, were purchased from afruit and vegetable market in Bari (Italy). Before treatment, individualfruits (apples and sweet cherries) and detached berries (table grapes)were randomized, surface sterilized in a 2% sodium hypochlorite solu-tion for 2 min and rinsed with running tap water for 1 min. Afterair-drying at room temperature, they were wounded (3 × 3 mmwound) with a sterile nail on the fruit equatorial area (two woundsper apple fruit; one wound per grape berry or cherry fruit) and keptat room temperature. After 30 min, 10 μl of conidial suspension wasapplied into each wound. For each isolate and fruit-type, tests wereconducted in triplicate and each replicate consisted of a tray containingsix fruits/berries. Replicates were individually wrapped into a plasticbag, avoiding contact between wounds and bags, and incubated at24 °C for 7 days. At the end of incubation, disease incidence (infectedwounds, %) and severity (lesion diameter, mm) were recorded. Forpatulin determination, rotted tissues were removed from the fruits ofeach replicate as suggested by Rychlik and Schieberle (2001), pooled,homogenized with a Sorvall Omnimixer (Sorvall Instruments, Norwalk,USA), and treated as reported by Sanzani et al. (2009a)with somemod-ifications. Briefly, an aliquot (5 g) of each homogenized tissues samplewas digested by 5 drops of a pectinase enzyme (1350–1650 U/g, OrsellS.R.L., Modena, Italy) in the presence of 5 ml of distilledwater overnightat room temperature. Sampleswere then centrifuged, filtered through aWhatman 4 paper followed by a 0.45 μm syringe filter, and analysed byHPLC as reported by Sanzani et al. (2009a). Samples containing patulinconcentrations outside the calibration rangewere appropriately dilutedwith mobile phase and re-analysed by HPLC. Results were expressed asμg of patulin per g of fresh weight.

2.5. Gene expression assay

The in vitro and in vivo growth experiments previously describedwererepeated to quantify the expression of genes coding the patulin biosyn-thetic enzymes isoepoxydon dehydrogenase (idh) and 6-methylsalycilicacid synthase (6msas) in the mycelium of each P. expansum isolategrown on its specific host (sweet cherries, table grapes and apples)and medium (CDA, GDA and ADA, respectively). For each isolate andhost/medium, tests were performed in triplicate (three sets of sixfruits/berries/dishes). By the end of incubation, total RNA was extractedas reported by Sanzani et al. (2009b, 2010) and DNAase treated as sug-gested by Promega (Wisconsin, USA). The RNA concentration, purityand integrity were determined by a spectrophotometer and by electro-phoresis ona 1.5% agarose gel in TAEbuffer (Sambrook andRussell, 2001).

Each RNA sample was reverse-transcribed using iScript cDNASynthesis Kit (Bio-Rad) and amplified in Real time PCR reactionsusing SYBR Green as fluorescent dye. The primer pairs, reaction mix,amplification conditions and standard curves were those reportedby Sanzani et al. (2009b). No template and non-reverse transcribedRNA controls were included in each run. Three technical replicateswere utilized for each gene and the average ratio of these valueswas used to determine the fold change in transcript level. Relativeexpression was evaluated by using the ΔΔCq method (Livak andSchmittgen, 2001). The calculation was made for each gene and setof strains (AI, CI and GI) on its host of isolation (and related growthmedium) considering the strains isolated from the other two hostsas calibrators. The relative expression values were automatically gen-erated by entering Cq values from housekeeping and target genes intoGene Expression Relative Quantification spreadsheets (Bio-Rad). Datawere transformed to log2 and levels of change were categorized as

501/489404331242190147

111/1106734

A

B

C

Fig. 1. Single nucleotide polymorphisms' (SNPs) presence and position (in red) (A); calibration curve for the primer pair used in High ResolutionMelting (HRM) analysis (B); primer pairspecificity confirmation by agarose gel electrophoresis and melt curve analysis (C).

115S.M. Sanzani et al. / International Journal of Food Microbiology 165 (2013) 111–120

follows: “low” ≥−1.0 to ≤1.0; “medium” ≥−2.0 to b−1.0 or >1.0to ≤2.0; “high” b−2.0 or >2.0 (Kim et al., 2008).

2.6. Data analyses

Data were subjected to ANOVA (one-way analysis of variance).Percentage data of incidence of decay were subjected to arcsine-square root transformation before ANOVA analysis. The data wereprocessed using the statistical software package Statistics for Windows(StatSoft, Tulsa, USA).

3. Results

3.1. Isolate collection and identification

Twenty-one putative P. expansum isolates (7 isolates/host type),identified by macro- and microscopic features, were selected amongstthose in the Penicillium culture collection for their host of isolationand extent of patulin production on PDA (Table 1, in bold). Their iden-tificationwas further confirmed by the amplification of the correspond-ing DNAwith P. expansum specific primers, which gave a single distinctproduct of 404 bp (data not shown), and by partial sequencing ofβ-tubulin gene. These two PCR tests enabled identification of all isolatesas P. expansum. The resulting 21 β-tubulin sequenceswere submitted toGenBank for accession numbers (Table 1) and aligned using specificprogrammes. The alignment revealed a certain variability amongstthem. In particular, single nucleotide polymorphisms (SNPs) weredetected at nucleotides 257 and 285 (Fig. 1A). They corresponded to atransversion (T to G) and a transition (C to T) mutation. Therefore,we decided to screen for the SNPs presence all the Penicillium spp.

collection, i.e. 54 P. expansum isolated from different hosts, 7P. crustosum and 5 P. chrysogenum (Table 1). To this purpose aprimer pair was designed on the sequence portion containing thetwo SNPs. The efficiency of this primer set was confirmed by build-ing up a calibration curve (Fig. 1B). A correlation coefficient of 0.997between DNA template dilutions (102–10−2 ng) and Cq valuesgiven by the instrument and an efficiency of 107.4% included inthe optimal range (90–110%) were recorded. Moreover, the meltcurve analysis and the agarose gel electrophoresis confirmed thespecificity of the reaction showing a single distinct amplificationproduct of 96 bp (Fig. 1C). These results allowed us to use the primerset in the following HRM analyses.

In these analyses P. crustosum and P. chrysogenum strains wereconsidered as negative controls. Furthermore, since no informationwas available at that time about the β-tubulin sequences of theother P. expansum strains, they were considered blind-tested for pres-ence of SNPs. Successful amplifications were achieved for all the iso-lates and, for each amplification product, normalized difference andderivative melt plots were produced (Fig. 2A, B). The HRM analysison normalized melting and difference curves always distributed the87 strains in six independent clusters, readily and consistently resolved(Fig. 2B). SNP markers allocated together most of the P. expansumstrains isolated from sweet cherries and table grapes (cluster 1, red),and in a separate cluster (cluster 2, blue) P. expansum strains originatingfrom apples. Furthermore, 4 more clusters were identified: cluster3 (dark green) that grouped P. chrysogenum strains coming fromsweet cherries, apricots, and almonds; cluster 4 (orange) that includedP. crustosum strains from pears and sweet cherries; cluster 5 (pink) thatcollected P. crustosum strains from almonds and raspberries; and cluster6 (bright green) which included P. crustosum originating from apples.

A B

C

Red = cluster 1Blue = cluster 2Dark green = cluster 3Orange = cluster 4Pink = cluster 5Bright green = cluster 6

Pex1

Pex59

Pex28

Pex84

Pex20

Pex44

Fig. 2. Discrimination of Penicillium expansum strains by HRM analysis. Normalized (A) and difference (B) plots obtained in HRM assays, performed using the couple of primersfor SNP markers on 87 Penicillium spp. isolates examined here. On each plot different colours indicate distinct profiles. RFU: Relative fluorescence units. Green and red columnsrepresent pre- and post-melting normalization regions. SNP group-based phylogenetic analysis of the representative P. expansum strains whose amplification product wasre-sequenced in order to confirm HRM results (C). (For interpretation of the references to colour in this figure legend, the reader is referred to the web of this article.)

Table 2In vitro growth (colony diameter, mm) on fruity media (apple dextrose agar, ADA; sweetcherry dextrose agar, CDA; table grape dextrose agar, GDA) and disease incidence(infected wounds, %) and severity (lesion diameter, mm) recorded in vivo on respectivefruits (apples, sweet cherries, and table grapes). Data were the mean of results fromseven strains isolated from apple (AI), sweet cherry (CI) and table grape (GI), each runin triplicate, ±standard error of the mean (SEM).

In vitro

Colony diameter (mm)

ADA CDA GDA

AI 50.6 ± 1.6 47.1 ± 1.9 48.5 ± 1.2CI 51.2 ± 2.3 54.1 ± 3.4 53.3 ± 2GI 50.3 ± 4.8 53.9 ± 1.9 55.3 ± 1.7

In vivo

Apple Cherry Grape

Incidence(%)

Severity(mm)

Incidence(%)

Severity(mm)

Incidence(%)

Severity(mm)

AI 62.1 ± 7 23.3 ± 4.4 51.2 ± 5.3 12 ± 8.3 53.6 ± 6.1 14.8 ± 1.9CI 58.1 ± 2.2 19.4 ± 3 76.3 ± 2.9 30.2 ± 5.9 70.4 ± 3.3 25.6 ± 2GI 54.5 ± 9.8 25.1 ± 1.3 63 ± 4.1 26.9 ± 2.7 81 ± 9.0 39.6 ± 6.4

116 S.M. Sanzani et al. / International Journal of Food Microbiology 165 (2013) 111–120

The clustering analysis was further confirmed by sequencing the ampli-fication product of one representative isolate per cluster. With theobtained sequences it was possible to obtain a phylogenetic tree inwhich the strains from each cluster were grouped separately (Fig. 2C).

3.2. Fungal growth and patulin production on different media and hosts

3.2.1. In vitroWhen isolates were grown on different artificial media (ADA, GDA

and CDA) made with their hosts of isolation (apples, table grapes, andsweet cherries, respectively), in general, the growth and sporulationrate were significantly (p b 0.05) lower when compared to thoseobserved on PDA (75 mm on average). However, generally no signif-icant differences in colony diameter were observed amongst thestrains from the three hosts grown on the same artificial medium, andamongst the strains originating from the same host on the three differ-entmedia (Table 2).Whereaswhen patulin productionwas considered,significant differences were recorded. In fact, considering the averageproduction on PDA (Table 1), the extent of toxin production provedto be influenced by the growth medium (Fig. 3A). In particular, onCDA, isolates from cherry produced the highest amount of patulin

0

100

200

300

400

500

600

700

PA

T(µ

g/g)

AI

CI

GI

0

20

40

60

80

100

120

140

160

180

200AI

CI

GI

AP

AT

(µg

/cm

2 )

B

In vitro

In vivo

ADAGDACDA

CHERRIES GRAPES APPLES

Fig. 3. Patulin production (μg/cm2 dish) by isolates of P. expansum cultured in vitro (A) ondifferent fruity growth media (CDA from sweet cherry, GDA from table grape, and ADAfrom apple) and in vivo (B) on sweet cherry, table grape, and apple fruits (μg/gfresh weight). Data represent the mean of 7 strains originating from apple (AI),7 from sweet cherry (CI) and 7 from table grape (GI), each run in triplicate, ±standarderror of mean (SEM).

0

2

4

6

8

10

12

14

16

18

206msas

idh

0

2

4

6

8

10

12

14

16

18

20

CI vs AI CI vs GI GI vs AI GI vs CI AI vs GI AI vs CI

CI vs AI CI vs GI GI vs AI GI vs CI AI vs GI AI vs CI

6msas

idh

Log 2(

rela

tive

expr

essi

on)

CDA

CHERRIES

Log 2(

rela

tive

expr

essi

on)

GDA

ADA

GRAPES

APPLES

A

B

Fig. 4. Relative expression (log2 transformed) of idh and 6-msas genes in the selectedtoxigenic P. expansum strains grown in vitro (A) on fruity media and in vivo (B) ontheir respective hosts. For each strain group on its respective host/medium data wereanalysed using the 2−ΔΔCqmethod, using the results of the other two groups as calibratorsand normalized using the β-tubulin housekeeping gene. Data represent the mean of 7strains originating from apple (AI), 7 from sweet cherry (CI) and 7 from table grape(GI), each run in triplicate, ±standard error of mean (SEM).

117S.M. Sanzani et al. / International Journal of Food Microbiology 165 (2013) 111–120

(52.76 μg/cm2) compared to isolates from apples and grapes; whereas,onGDA the isolates fromgrapewere the best producers (167.75 μg/cm2);and finally, on ADA the highest amount of patulin was produced by appleisolates (11.58 μg/cm2). However, overall, GDAandADAwere the growthmedia on which the highest and lowest amounts of patulin were pro-duced, respectively (Fig. 3A).

3.2.2. In vivoWhen P. expansum strains were grown on the corresponding fresh

fruits (i.e. apples, sweet cherries, and table grapes), significant differ-ences in disease incidence and severity were observed when theoriginal the host was taken into account (Table 2). In particular, thehighest disease incidence and severity were caused by sweet cherry(76.3% and 30.2 mm), table grape (81.0% and 39.6 mm) and apple(62.1% and 23.3 mm) strains on their respective host of isolation(Table 2). Similarly, evaluating the average patulin accumulation byCI, GI and AI strains, each group proved to be the strongest patulinproducer on its related host of isolation (sweet cherries, table grapesand apples), producing 500.17, 554.19 and 93.05 μg/g, respectively(Fig. 3B). On the whole, these results show that the highest decayincidence/severity and patulin accumulation were observed on tablegrape, whereas apple was the host in which the decay was less ag-gressive (Table 2) and the lowest amount of patulin was produced(Fig. 3B).

3.3. Gene expression assay

The relative expression of 6msas and idh genes was evaluated in themycelium of the selected 21 strains grown on apples, sweet cherriesand table grapes and their respective artificial media (Fig. 4). The rela-tive expression calculation was made for each set of strains on its hostof isolation (and related growth medium) considering the strains orig-inating from the other two hosts as calibrators.

3.3.1. In vitro (Fig. 4A)On CDA the 6msas expression extent in the mycelium of sweet

cherry strains was 7 and 3.6-fold higher than that of apple and tablegrape strains, respectively; whereas, idh expression was 2.9 and2.1-fold higher than that recorded for apple and table grape strains,respectively. On GDA the 6msas expression in the mycelium of tablegrape isolates was 13.8 and 12.7-fold higher than that recorded forapple and sweet cherry strains, respectively; whereas, idh expressionwas 9.9 and 8.2-fold higher than that of apple and sweet cherrystrains, respectively. Finally, on ADA 6msas expression in the myceliumof apple strains was 2.3 and 3.6-fold higher than that of table grape andsweet cherry strains, respectively; whereas, idh expression was 6.5 and4.3-fold higher than that of table grape and sweet cherry strains, respec-tively. In general, the highest induction of the two genes under evalua-tion was observed on GDA in the mycelium of the strains isolated fromtable grape, followed by sweet cherry and apple strains, respectively.

118 S.M. Sanzani et al. / International Journal of Food Microbiology 165 (2013) 111–120

3.3.2. In vivo (Fig. 4B)On sweet cherries the 6msas expression extent in the mycelium of

its strains was 11.2 and 5.8-fold higher than that of apple and tablegrape strains, respectively; whereas, idh expression was 3.4 and3.0-fold higher than that recorded for apple and table grape strains,respectively. On table grape the 6msas expression in the myceliumof its strains was 17 and 14.2-fold higher than that recorded forapple and sweet cherry strains, respectively; whereas, idh expressionwas 12.1 and 9.0-fold higher than that of apple and sweet cherrystrains, respectively. Finally, on apple 6msas expression in the myce-lium of its isolates was 2.9 and 4.4-fold higher than that of tablegrape and sweet cherry strains, respectively; whereas, idh expressionwas 7.9 and 5-fold higher than that of table grape and sweet cherrystrains, respectively.

In conclusion, the highest induction of the two genes under eval-uation was observed in the mycelium of table grape strains grownon their host of isolation.

4. Discussion

Although bluemould of apples ismainly associatedwith P. expansum,several other species have also been reported to be responsible for therot (Pianzzola et al., 2004). Therefore, a precise characterization of thecausal agent at a species level is important, especially for those fruits des-tined for the fruit processing industry, because of the putative presenceof patulin. However, Penicillium identification is not easy. It is a largegenus, and many species have very similar macro- and microscopicfeatures. At the same time a great deal of variability within the speciesis reported, therefore, their unambiguous identification often requiresother approaches such as secondary metabolite production profile(Frisvad and Samson, 2004) and molecular identification (Guercheet al., 2004). In the present investigation patulin production ability, spe-cies specific primers (Marek et al., 2003) and β-tubulin gene sequencing(Frisvad and Samson, 2004; Glass and Donaldson, 1995) were usedto confirm the identity of 21 Penicillium spp. isolates selected for thetrials. They permitted the unequivocal identification of the strains asP. expansum and patulin producers (0.6–38.4 μg/cm2). Furthermore,the alignment of the β-tubulin sequence fragments obtained showedevidence of a certain degree of variability amongst isolates. Indeed twoSNPs were found, that encouraged us to screen the entire Penicilliumcollection available at the Department of Soil, Plant and Food Sciencesof the University of Bari by means of HRM.

In the present study we report for the first time, as far as we know,the successful use of HRM on Penicillium spp. associated with bluemould of pome fruits. A total of 87 Penicillium isolates belonging tothree different species and isolated from several hosts were testedaccording to the presence of the two above mentioned SNPs. Geneticmutations may comprise single/multiple nucleotide change, insertionand/or deletion. SNPs have been categorized in four classes: I) C>T,T>C, G>A, A>G; II) CbA, A>C, G>T, T>G; III) C>G, G>C; IV) A>T,T>A. The class IV is the rarest and most difficult to identify since thedifference is just a hydrogen bond. The mutations recorded (T to Gand C to T) amongst the isolates under examination belonged toclass I and II which are the commonest. These SNPs generated uniqueHRM profiles able to group together and exclusively all the strainsbelonging to the same species (i.e. P. expansum, P. chrysogenum, andP. crustosum). Furthermore, interestingly, within each species thestrains were further subdivided in clusters, grouping together strainsisolated from the same host. These findings seem to support thehypothesis of a sort of host specialization amongst the isolates of asame species. Furthermore, as supporting circumstance it has to beconsidered that the growing season and/or geographic localizationof the three hosts does not overlap (Carter et al., 2000).

A high degree of specificity has been observed in many plant–pathogen interactions (Peever et al., 1999). For instance, Carter et al.(2000) found that Fusarium graminearum isolates from different

hosts were not evenly distributed between the groups identified bymolecular approaches, thus indicating that an element of host prefer-ence may be involved. Indeed, group B isolates were recovered lessfrequently from wheat and rice than from maize, suggesting thatthese isolates might be less aggressive to wheat and rice than tomaize. Although further research on Penicillium genus with otherphylogenetically informative markers and a larger sample of well-described species from a wider host range is required, these pre-liminary results were particularly encouraging for the pursuance ofthe study.

Research on the genetics of host specificity focussed mainly onintraspecific host variation (race or cultivar specificity) and species-specificity (ability of one species of pathogen to cause disease onone or more host species or genera) (Peever et al., 1999). In patho-gens with a sexual stage that can be induced readily in the laboratory,host specificity can be studied directly by following the segregation ofpathogenicity on one or more hosts amongst the progeny of crosses(Sweigard et al., 1995). However, in pathogens that reproduce asexu-ally (in Penicillium genus sexual reproduction is rare and documentedjust in very few species) or that are not amenable to genetic analyses,host-associated differences in neutral molecular markers can be in-vestigated in population samples of isolates from different hosts(Freeman et al., 1996). In our study, the growth and patulin produc-tion extent of 21 representative P. expansum strains isolated fromdifferent hosts (sweet cherries, table grapes and apples) were testedin vivo on the three hosts themselves and in vitro on their respectiveartificial growth media. Particular attention was devoted to the be-haviour of each strain group on the different hosts/media in relationto its host of isolation.

All P. expansum isolates tested in vivo proved to be able to cause bluemould in hosts (table grapes and sweet cherries), which are less com-monly associated with the disease than apples. However, some sig-nificant differences were observed in their pathogenicity/virulence.Indeed, when the isolates were re-grown on their host of isolation,they proved to be more aggressive and to produce more patulin com-pared to the groups of isolates coming from the other two hosts underevaluation. On average, apple proved to be the least susceptible fruitto the tested P. expansum groups, and the fruit in which the lowestamount of patulin was recorded. These findings further confirm thepotential of fruit types, considered less common hosts for P. expansum,to support patulin occurrence at a similar or even higher extent, alreadyreported in other studies (Larsen et al., 1998; Neri et al., 2010; Reddyet al., 2010). Thus they might represent a relevant source of patulinintake by consumers, especially children.

The highest incidence and severity of infection was in general ob-served in fruits with softer texture, such as sweet cherries but aboveall table grapes, compared to firmer ones such as apples. Since anincrease of pectin depolymerization also occurs during P. expansuminfection (Miedes and Lorences, 2006), colonization of tissues wasprobably facilitated in softer fruits where the pathogen could alreadyfind favourable conditions for the progress of infection. Furthermore,the results of patulin production obtained in vivo are consistent withthose obtained in vitro, whereas a discrepancy was observed with re-sults of in vitro growth, where the differences observed amongst thethree groups of isolates on the three fruit-based media proved notto be significant. This difference between in vitro and in vivo resultsis in agreement with previous assays on P. expansum growth andpatulin production by Neri et al. (2010) and has several possibleexplanations. For instance, it might be related to host constituentsor signal molecules that do not survive to the sterilization proceduresused during preparation of artificial growth medium or that may bereleased only by a living host. Moreover, it could be related to patho-genicity factors produced by the fungus itself in presence of a partic-ular host as in the case of Alternaria citri. Indeed, the specificity ofA. citri for rough lemon (rough lemon pathotype) or Dancy tangerine(tangerine pathotype) appeared to be due to the production of

119S.M. Sanzani et al. / International Journal of Food Microbiology 165 (2013) 111–120

host-specific toxins (Kohmoto et al., 1979), and sensitivity to thetoxin was strongly correlated to the host specificity of the fungal iso-lates producing the toxin (Kohmoto et al., 1991). In this regard, therole played by patulin should be taken into consideration. In our in-vestigation, patulin production seemed to be correlated with fungalaggressiveness observed in vivo on the different hosts, since for eachhost type, the isolate group producing the lowest amount of patulinproved to be the least pathogenic/virulent. Furthermore, recently ithas been found by Sanzani et al. (2012) that P. expansum mutants,in which the key patulin biosynthetic gene 6msas had been disrupted,were less pathogenic/virulent compared to their wild-type and thatthey regained their aggressiveness once the patulin was exogenouslyapplied.

How patulin may exert its effect is still not completely clear.During the course of invasion, plant pathogens encounter severalhost defensive strategies, including the production of reactive oxygenspecies (ROS), primarily superoxide (O2

−) and hydrogen peroxide(H2O2), against which pathogens produce antioxidant enzymes(e.g. catalase and superoxide dismutase). In a study by Qin et al.(2007), upon exposure of P. expansum to the antifungal salt borate,its antioxidant enzymes exhibited reduced levels of expression(with consequent accumulation of ROS and carbonylated proteins)and the analysis of its extracellular proteome revealed down-regulation of polygalacturonase. Oxidative stress, meanwhile, hasbeen reported to trigger mycotoxin biosynthesis. For instance, Reverberiet al. (2005, 2006, 2008) demonstrated that in A. parasiticus culturedunder conducive conditions for toxin production, both the oxidativeburst and the activity of antioxidant enzymes was much higher than innon-conducivemedia. Thus, aflatoxinswould help tomaintain the oxida-tive status at levels less harmful to the fungus, besides providing otherecological advantages (Reverberi et al., 2010). Although the transductionpathways related to patulin regulation are still to be completely discov-ered, the evidence of the control exerted by oxidative stress (Sanzaniet al., 2009b; Tolaini et al., 2010) suggests the existence of regulatorymechanisms similar to those acting in Aspergillus.

In our study the influence of the host/medium on the patulin pro-duction potential of the isolates under investigation was analysedeven from a transcriptomic point of view. Indeed, the expression of6msas, but also of the other patulin biosynthetic gene idh, was evalu-ated in each host/medium for each strain group compared to theothers. The results highlighted an up-regulation of the two genes,particularly 6msas, when the strains were grown on their particularhost of isolation or on the corresponding medium. The highestup-regulation was observed for table grape strains, particularly inin vivo trials. These results are consistent with the patulin quantifica-tion data, as well as with the in vivo growth tests, in which the tablegrape isolates proved the most aggressive. Furthermore, the promi-nent role of the living host seemed to be additionally supported bythe highest up-regulation of patulin biosynthetic pathway genes ob-served in vivo as compared to in vitro trials. Finally, an inverse behav-iour of idh to 6msaswas observed on apples/ADA as compared to thaton sweet cherry/CDA or table grape/GDA, with idh being up-regulatedto a higher extent than 6msas. This finding is in agreement with thelowest patulin production observed on apple/ADA and the HRM re-sults in which apple isolates grouped separately from those fromsweet cherry and table grape, which grouped together. The 6mascatalyzes the first committed step of patulin biosynthetic pathwayand thus its lowest up-regulation, possibly due to response of appletissues and/or composition, might account for the lowest patulinaccumulation recorded compared to other tissues/media. In a paperby Abdel-Hadi et al. (2011) a decrease in mRNA expression of afla-toxin biosynthetic gene aflD, because of silencing by siRNA, resultedin an accumulation of intermediate compounds and a decrease inaflatoxin B1 production. Authors reported also aflD ability to influ-ence the expression of the aflatoxin regulatory gene aflR, since,following aflD knockdown, a down-regulation of aflR was observed.

Although the patulin biosynthetic pathway has been less extensivelystudied than that of aflatoxin, a parallelism between 6msas/idhbehaviour on apple/ADA could not be excluded, although, as far aswe know, neither of the two genes are reported as transcriptionactivator.

In conclusion the fruit host seems to influence patulin production ata transcriptomic level, and P. expansum strains seem to benefit in theiraggressiveness by the increased quantity of this putative pathogenic/virulence factor. A variation at a genetic level amongst isolates of thesame species observed by HRM analyses is particularly interesting andseems to confirm the hypothesis of the existence of host specificity,although this needs to be further confirmed. HRM is a low-cost identifi-cation method compared to other diagnostic assays because no expen-sive or specialized reagents are needed. Hence, P. expansum-specificHRM-based protocol developed here could be used to facilitate andpromote high-throughput screenings aimed at revealing important ep-idemiological data and gain further insight into trends of this species,such as host range, temporal or geographical tendencies and otherfundamental observations related to pathogenicity/virulence. More-over, the obtained information on P. expansum–host interaction mightbe useful for the management of blue mould, helping to correctly cali-brate preventive disease control measures, as the few permitted fungi-cides are rapidly becoming ineffective and legislation concerningmycotoxins and fungicides is becoming increasingly restrictive.

References

Abdel-Hadi, A.M., Caley, D.P., Carter, D.R.F., Magan, N., 2011. Control of aflatoxinproduction of Aspergillus flavus and Aspergillus parasiticus using RNA silencingtechnology by targeting aflD (nor-1) gene. Toxins 3, 647–659.

Carter, J.P., Rezanoor, H.N., Desjardin, A.E., Nicholson, P., 2000. Variation in Fusariumgraminearum isolates fromNepal associated with their host of origin. Plant Pathology49, 452–460.

Castoria, R., Mannina, L., Durán-Patrón, R., Maffei, F., Sobolev, A.P., De Felice, D.V.,Pinedo-Rivilla, C., Ritieni, A., Ferracane, R., Wright, S.A.I., 2012. Conversion of themycotoxin patulin to the less toxic desoxypatulinic acid by the biocontrol yeastRhodosporidium kratochvilovae strain LS11. Journal of Agricultural and FoodChemistry59, 11571–11578.

European Commission, 2006. Commission Regulation (EC) No. 1881/2006 settingmaximum levels for certain contaminants in foodstuffs. Official Journal of the EuropeanUnion 364, 5–24.

Freeman, S., Katan, T., Shabi, E., 1996. Characterization of Colletotrichum gloeosporioidesisolates from avocado and almond fruits with molecular and pathogenicity tests.Applied and Environmental Microbiology 62, 1014–1020.

Frisvad, J.C., Samson, R.A., 2004. Polyphasic taxonomy of Penicillium subgenus PenicilliumA guide to identification of food and air-borne terverticillate Penicillia and theirmycotoxins. Studies in Mycology 49, 1–174.

Garritano, S., Gemignani, F., Voegele, C., Nguyen-Dumont, T., Le Calvez-Kelm, F., DeSilva, D., Lesueur, F., Landi, S., Tavtigian, S.V., 2009. Determining the effectivenessof High Resolution Melting analysis for SNP genotyping and mutation scanningat the TP53 locus. BMC Genetics 10, 5.

Glass, N.L., Donaldson, G.C., 1995. Development of primer sets designed for use withthe PCR to amplify conserved genes from filamentous Ascomycetes. Applied andEnvironmental Microbiology 61, 1323–1330.

Gori, A., Cerboneschi, M., Tegli, S., 2012. High-Resolution Melting Analysis as a powerfultool to discriminate and genotype Pseudomonas savastanoi pathovars and strains.PLoS ONE 7, e30199.

Guerche, S.L., Garcia, C., Darriet, P., Dubourdieu, D., Labarere, J., 2004. Characterization ofPenicillium species isolated from grape berries by their internal transcribed spacer(ITS) sequences and by gas chromatography-mass spectrometry analysis of geosminproduction. Current Microbiology 48, 405–411.

Howlett, B.J., 2006. Secondary metabolite toxins and nutrition of plant pathogenicfungi. Current Opinion in Plant Biology 9, 371–375.

Kim, J.H., Yu, J., Mahoney, N., Chan, K.L., Molyneux, R.J., Varga, J., Bhatnagar, D.,Cleveland, T.E., Nierman, W.C., Campbell, B.C., 2008. Elucidation of the functionalgenomics of antioxidant-based inhibition of aflatoxin biosynthesis. InternationalJournal of Food Microbiology 122, 49–60.

Kohmoto, K., Scheffer, R.P., Whiteside, J.O., 1979. Host-selective toxins from Alternariacitri. Phytopathology 69, 667–671.

Kohmoto, K., Akimitsu, K., Otani, H., 1991. Correlation of resistance and susceptibility ofcitrus to Alternaria alternatawith sensitivity to host specific toxins. Phytopathology81, 719–722.

Larsen, T.O., Frisvad, J.C., Ravn, G., Skaaning, T., 1998. Mycotoxin production by Penicilliumexpansum on blackcurrant and cherry juice. Food Additives and Contaminants 15,671–675.

Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCT method. Methods 25, 402–408.

120 S.M. Sanzani et al. / International Journal of Food Microbiology 165 (2013) 111–120

Maggio-Hall, L.A., Wilson, R.A., Keller, N.P., 2005. Fundamental contribution of beta-oxidation to polyketide mycotoxin production in planta. Molecular Plant MicrobeInteraction 18, 783–793.

Marek, P., Annamalai, T., Venkitanarayanan, K., 2003. Detection of Penicillium expansumby polymerase chain reaction. International Journal of Food Microbiology 89,139–144.

Mari, M., Neri, F., Bertolini, P., 2009. Management of important diseases in Mediterraneanhigh value crops. Stewart Postharvest Review 2, 2.

Miedes, E., Lorences, E.P., 2006. Changes in cell wall pectin and pectinase activityof apple and tomato fruits during Penicillium expansum infection. Journal of theScience of Food and Agriculture 86, 1359–1364.

Neri, F., Donati, I., Veronesi, F., Mazzoni, D., Mari, M., 2010. Evaluation of Penicilliumexpansum isolates for aggressiveness, growth and patulin accumulation in usualand less common fruit hosts. International Journal of Food Microbiology 143,109–117.

Peever, T.L., Canihos, Y., Olsen, L., Ibañez, A., Liu, Y.-C., Timmer, L.W., 1999. Populationgenetic structure and host specificity of Alternaria spp. causing brown spot ofMinneola Tangelo and Rough Lemon in Florida. Phytopathology 89, 851–860.

Pianzzola, M.J., Moscatelli, M., Vero, S., 2004. Characterization of P. expansum isolatesassociated with Blue Mold on apple in Uruguay. Plant Disease 88, 23–28.

Piemontese, L., Solfrizzo, M., Visconti, A., 2005. Occurrence of patulin in conventionaland organic fruit products in Italy and subsequent exposure assessment. FoodAdditives and Contaminants 22, 437–442.

Pitt, J.I., Hocking, A.D., 2009. Fungi and Food Spoilage, Third edition. Springer, NewYork, USA.Pitt, J.I., Spotts, R.A., Holmes, R.J., Cruickshank, R.H., 1991. Penicillium solitum revived

and its role as a pathogen of pomaceous fruit. Phytopathology 81, 1108–1112.Qin, G., Tian, S., Chan, Z., Li, B., 2007. Crucial role of antioxidant proteins and hydrolytic

enzymes in pathogenicity of Penicillium expansum: analysis based on proteomicsapproach. Molecular & Cellular Proteomics 6, 425–438.

Reddy, K.R.N., Spadaro, D., Lore, A., Gullino, M.L., Garibaldi, A., 2010. Potential of patulinproduction by Penicillium expansum strains on various fruits. Mycotoxin Research26, 257–265.

Reverberi, M., Fabbri, A.A., Zjalic, S., Ricelli, A., Punelli, F., Fanelli, C., 2005. Antioxidantenzymes stimulation in Aspergillus parasiticus by Lentinula edodes inhibits aflatoxinproduction. Applied Microbiology and Biotechnology 69, 207–215.

Reverberi,M., Zjalic, S., Ricelli, A., Fabbri, A.A., Fanelli, C., 2006. Oxidant/antioxidant balancein Aspergillus parasiticus affects aflatoxin biosynthesis. Mycotoxin Research 22, 39–47.

Reverberi, M., Zjalic, S., Ricelli, A., Punelli, F., Camera, E., Fabbri, C., Picardo, M., Fanelli,C., Fabbri, A.A., 2008. Modulation of antioxidant defense in Aspergillus parasiticus

is involved in aflatoxin biosynthesis: a role for the ApyapA gene. Eukaryotic Cell 7,988–1000.

Reverberi, M., Ricelli, A., Zjalic, S., Fabbri, A.A., Fanelli, C., 2010. Natural functions ofmycotoxins and control of their biosynthesis in fungi. Applied Microbiology andBiotechnology 87, 899–911.

Rychlik, M., Schieberle, P., 2001. Model studies on the diffusion behavior of themycotoxin patulin in apples, tomatoes, and wheat bread. European Food Researchand Technology 212, 274–278.

Sambrook, J., Russell, D.W., 2001. Molecular Cloning: a laboratory Manual, 3rd ed. ColdSpring Harbor Laboratory Press, New York, USA.

Samson, R.A., Pitt, J.I., 2000. Integration of Modern Taxonomic Methods for Penicilliumand Aspergillus Classification. Harwood Academic Pulishers, London, UK.

Sanderson, P.G., Spotts, R.A., 1995. Postharvest decay of winter pear and apple fruitcaused by species of Penicillium. Phytopathology 85, 103–110.

Sant'Ana, A.d.S., Rosenthal, A., De Massaguer, P.R., 2008. The fate of patulin in applejuice processing: a review. Food Research International 41, 441–453.

Sanzani, S.M., De Girolamo, A., Schena, L., Solfrizzo, M., Ippolito, A., Visconti, A., 2009a.Control of Penicillium expansum and patulin accumulation on apples by quercetinand umbelliferone. European Food Research and Technology 228, 381–389.

Sanzani, S.M., Schena, L., De Girolamo, A., Nigro, F., Ippolito, A., 2009b. Effect of quercetinand umbelliferone on the transcript level of Penicillium expansum genes involved inpatulin biosynthesis. European Journal of Plant Pathology 125, 223–233.

Sanzani, S.M., Schena, L., De Girolamo, A., Ippolito, A., González-Candelas, L., 2010.Characterization of genes associated to induced resistance against Penicilliumexpansum in apple fruit treated with quercetin. Postharvest Biology and Technology56, 1–11.

Sanzani, S.M., Reverberi, M., Punelli, M., Ippolito, A., Fanelli, C., 2012. Study on the roleof patulin on pathogenicity and virulence of Penicillium expansum. InternationalJournal of Food Microbiology 153, 323–331.

Schena, L., Nigro, F., Ippolito, A., 2002. Identification and detection of Rosellinia necatrixby conventional and real-time Scorpion-PCR. European Journal of Plant Pathology108, 355–366.

Sweigard, J.A., Carroll, A.M., Kang, S., Farrall, L., Chumley, F.G., Valent, B., 1995. Identifi-cation, cloning, and characterization of PWL2, a gene for host species specificity inthe rice blast fungus. The Plant Cell 7, 1221–1233.

Tolaini, V., Zjalic, S., Reverberi, M., Fanelli, C., Fabbri, A.A., Ricelli, A., 2010. Lentinulaedodes enhances the biocontrol activity of Crypyptococcus laurentii againstPenicillium expansum contamination and patulin production in apple fruits.International Journal of Food Microbiology 138, 243–248.