Biodiversity and ITS-RFLP Characterisation of Aspergillus Section Nigri Isolates in Grapes from Four Traditional Grape-Producing Areas in Greece

A study on the occurrence of Aspergillus section Nigri species on grapes from four traditional grape-producing areas in Greece during the 2011/2012 vintage, and their capability to produce OTA was conducted. One hundred and twenty-eight black aspergilli isolates were characterised at the species level initially by the use of morphological criteria in accordance with appropriate keys, followed by molecular characterisation performed with Polymerase Chain Reaction–Restriction Fragment Length Polymorphism (PCR-RFLP) of the 5.8 ribosomal RNA gene Internal Transcribed Spacer region (5.8 rRNA ITS). Restriction enzyme digestion of the ITS amplicons using the HhaI, HinfI and RsaI, endonucleases distinguished eleven different patterns of restriction fragment length polymorphism (RFLP), four for each of the HhaI and RsaI digests and three for HinfI. From a total number of 128 individual isolates, 124 were classified into four Aspergillus species corresponding to A. carbonarius, A. tubingensis, A. japonicus and A. ibericus, and the remaining 4 were classified as members of the A. niger aggregate. A. carbonarius and A. tubingensis being the main representative species were equally counted, with higher geographical representation of the former in southern and the latter in northern regions, respectively. All isolates were tested for their ochratoxigenic potential by use of High Performance Liquid Chromatography (HPLC) and Enzyme Linked Immuno Sorbent Assay (ELISA), resulting in significant interspecies differences in OTA production.

OTA is receiving attention due to its high incidence in a wide range of food commodities such as cereal-based products, coffee, spices, nuts, olives and grape-derived products [8,9,10,11,12,13]. RASFF Annual Report places OTA as the mycotoxin with most notifications for fruit and vegetables in 2012, while is being always second after aflatoxins in notifications on mycotoxins for food and feed for over the last decade [14]. Several publications report the high occurrence of OTA in wine and grape products from different European countries, with higher OTA levels reported for products originating from southern European regions and in particular in regions with a Mediterranean climate [15,16,17,18,19,20]. Since 2005 the European Commission has imposed regulatory limits for OTA, and has established a 2 mg l 21 maximum level of OTA in wine and grape products [21,22].
Several studies performed worldwide have shown that OTA is produced during infection of grapes in vineyards mainly by mycotoxigenic strains of black aspergilli (section Nigri), in particular Aspergillus carbonarius and species belonging to the Aspergillus niger aggregate such as Aspergillus niger and Aspergillus tubingensis, [18,19,20,23,24,25]. A. carbonarius although less common than other black aspergilli, is considered to be the predominant species responsible for OTA contamination in grapes and wine, because of the ability of almost all of its strains to produce high levels of the toxin [26,27]. Due to the variations in the toxin production potency of different aspergilli, it is of great importance to identify the Aspergillus species accurately in order to define potential toxicological risks at an early stage [28].
Black aspergilli are difficult to classify following only morphological criteria [29]. Though A. carbonarius can be easily recognized, closely related morphospecies like the biseriate species included in the A. niger aggregate, or uniseriate species such as Aspergillus aculeatus and Aspergillus japonicus, have been always difficult to distinguish. Among the molecular approaches used to decipher the Aspergillus taxonomy [30,31,32,33,34,35,36,37,38,39], Polymerase Chain Reaction -Restriction Fragment Length Polymorphism (PCR-RFLP) has been used successfully to identify Aspergillus species [33,39]. The non-coding Internal Transcribed Spacer regions (ITS) due to their high degree of sequence polymorphism have been used widely for fungi molecular systematics at the species level [40,41,42]. Standard primer pairs have been used for selective amplification of these fungal sequences and have been proved useful for the identification of Aspergillus species [33,39].
The objective of this study was to characterize (i) the biodiversity of black aspergilli isolates from vineyards of four traditional grape producing regions in Greece during the 2011/2012 vintage using molecular techniques and (ii) study their ochratoxigenic potential. To our knowledge this is the first study with the objective to monitor the population of black aspergilli by molecular means in Greece.

Ethics Statement
Part of this study was carried out in private lands. No specific permissions were required for these locations/activities that took place and all the land owners agreed to perform this research in their site. The field studies did not involve endangered or protected species.

Sampling and culturing conditions
Thirty-three vineyards, located at four traditional grapeproducing regions of Greece, were chosen for sampling during the 2012 harvesting period (from August to October). The sampled vineyards were located in Iraklion prefecture of Crete (n = 8), Mesogeia province of Attica (n = 11), Corinthia and Arcadia prefectures of Peloponnese (n = 7), Thessaloniki, Imathia, Florina and Pella prefectures of Macedonia (n = 7). Selection of sampling areas was made with provision to represent the whole range of the typical climatic conditions of Greece, i.e. the Peloponnese from Southern Greece and Attica from Central Greece, having both a landlocked Mediterranean climate with vineyards of high and low altitudes, respectively, Crete from the Southern Aegean Sea with insular Mediterranean climate, and Macedonia from Northern Greece having a typical mountainous climate with vineyards of high altitude.
From every vineyard five plants were selected along two major diagonal transects and two bunches were collected from each plant, resulting in a total of 10 bunches per sampled vineyard. Bunches were kept separate in sterile plastic bags and stored in portable refrigerators during transfer to the laboratory where the analysis took place within 24 hours from sampling.
From each bunch five healthy berries were randomly selected and placed directly on the surface of Dichloran Rose Bengal Chloramphenicol (DRBC) medium (LabM, UK). The plates were incubated in the dark at 25uC for 7 days.

Mycoflora enumeration and black aspergilli isolation and identification
After incubation, incidence of infected with black aspergilli berries and distribution of most dominant grape mycoflora at genus level, were recorded. A representative number from the black aspergilli from every vineyard were isolated from DRBC and sub-cultured on Malt Extract Agar (MEA, LabM) and Czapek Dox Agar (CD, Oxoid) for further identification at species level. The initial identification of the different strains of Aspergillus section Nigri was performed using macroscopic and microscopic morphological criteria in accordance with appropriate keys [29,43,44]. The reference strains of A. carbonarius, A. niger, A. tubingensis, A. ochraceus and A. westerdijkiae were kindly provided by Prof. N. Magan from the Mycology Group of Cranfield University. The isolates were preserved at 280uC in the culture collection of the Department of Food Science and Human Nutrition of the Agricultural University of Athens, Greece.

DNA extraction
For DNA extraction all strains were grown in Yeast Extract Sucrose broth (YES; contained per litre 20 g yeast extract and 150 g sucrose) at 30uC for 2 days. Mycelia were collected, washed briefly with ethanol 96%, and dried using Whatman No. 1 filter paper. Approximately 200 mg of mycelia from each individual strain were frozen in liquid nitrogen and ground to a fine powder. DNA extractions were performed using the NucleoSpin Plant DNA kit (Macherey-Nagel, Germany) according to the manufacturer's instructions.

PCR reactions and Restriction Endonuclease DNA digestions
The 5.8 S-ITS region was amplified by PCR using universal primers ITS 1 and ITS 4 [45]. PCR reactions were performed in a 50 mL final volume, containing 16 standard reaction buffer (New England Biolabs, UK), 2.0 mM MgCl 2 , 300 mM dNTPs (each), 300 nM primers (each), 100 ng DNA template and 1.25 U of Taq DNA polymerase (New England Biolabs, UK). The PCR reactions were performed in a MJ Research PTC-200 thermal cycler (Bio-Rad Laboratories, USA), starting with an initial denaturation step at 95uC for 5 min, followed by 37 cycles consisting of 30 sec at 95uC, 30 sec at 52uC and 40 sec at 72uC, and a final extension step at 72uC for 10 min. PCR products were digested with the HhaI, HinfI and RsaI (New England Biolabs, UK) restriction endonucleases. Digestions were performed at 37uC for 3 h, in a 20 mL reaction volume containing 2 mL of 10X reaction buffer, 10 mL amplicon, 1.5 U restriction endonuclease, and 0.2 mL BSA (10 mg mL 21 ) for HhaI digestions. PCR amplicons and their restriction digestion fragments were separated by electrophoresis at 100 V, 16TAE buffer, in 1% and 3% agarose gels, respectively. Agarose gels were subsequently stained in ethidium bromide solution (0.5 mg ml 21 ), and DNA bands were visualized under Ultra Violet (UV) light using a Gel Doc XR + system (Bio-Rad Laboratories, USA). Molecular sizes were estimated by comparison with the DNA standard GeneRuler 50 bp and 100 bp DNA ladders (Thermo Scientific, USA).

Sequencing and Phylogenetic Analysis
PCR products were purified using a QIAquick PCR purification kit (Qiagen, USA). Sequencing was performed for both strands using primers ITS1 and ITS4 with the BigDye Terminator v3.1 cycle sequencing kit (Life Technologies, USA) in an ABI3730 xl Genetic Analyzer (Applied Biosystems, Life Technologies, USA) automatic DNA sequencer (Cemia, Greece).
Alignment of the 5.8S-ITS region sequences was performed using the CLUSTALV multiple sequence alignment program available at EBI (EMBL, UK  [46]. The NJ tree and the statistical confidence of a particular group of sequences in the tree, evaluated by bootstrap test (1000 pseudoreplicates), were performed using the computer program MEGA version 3.0 [47].

OTA extraction
All 128 black aspergilli isolates were centre inoculated to CYA medium (which contained per litre: K 2 HPO, 1 g; Czapek concentrate, 10 ml; trace metal solution, 1 ml; yeast extract, 5 g; sucrose, 30 g; agar, 15 g) [41] and incubated at 25uC for 7 days in order to assess their ochratoxigenic potential. Aspergillus section Nigri isolates was subjected to Ochratoxin A determination according to the Bragulat et al. protocol [48] with a slight modification in the OTA extraction step. Instead of removing 3 small agar plugs, the whole content of the Petri dish was removed and used for OTA extraction [49,50]. Specifically, the content of the Petri dish (substrate and mycelium) was weighted in order to express OTA production per g of substrate and extraction took place with 100 ml of an 80/20 methanol/water solution of HPLC grade purity when prepared for HPLC analysis and a 50/50 solution when prepared for the ELISA method. The weighted substrates were blended with the solutions for 2 min and left for a total of 30 min before filtered, at first through a Whatman No 1 filter paper, and subsequently through Millex nylon membrane filter of 0.2 mm pore size (EMD Millipore Corp. Billerica, USA), and kept at 280uC until analysis. Additionally, known concentrations of OTA (50, 100 & 500 ppb) were spiked on CYA and recovery rates for both HPLC and ELISA methods were estimated, resulting in satisfactory recovery percentages of 96-99% and 84-95% for the former and latter method, respectively.

HPLC Analysis
Ochratoxin A analysis was performed using reverse-phase High Performance Liquid Chromatography with fluorometric detection (HPLC-FLD). This consisted of a JASCO AS-2055Plus autosampler, a JASCO LC-Net II/ADC system controller, a JASCO PU-980/LC-980-02 pump, and a JASCO FP-2020Plus fluorescence detector (JASCO Inc., Easton, USA). The samples were separated using a C18 analytical column (25064.6 nm, 4 mm, Resteck Co., Pinnacle II, Bellefonte, USA) under isocratic conditions at a flow rate of 1 ml min 21 of the mobile phase (water/acetonitrile/acetic acid: 99/99/2). All chemicals used for HPLC analysis were HPLC grade (methanol and acetic acid: Sigma-Aldrich Co., Germany; water and acetonitrile: Carlo Erba Reactifs SDS, Val de Ruill, France). An excitation wavelength of 333 nm and an emission wavelength at 460 nm were used for UV detection. Standard solutions were made from stock ochratoxin A solution (10.06 mg ml 21 in acetonitrile; Biopure, Romer Labs Diagnostics GmbH, Tulln, Austria) in mobile phase and a recovery study took place by spiking known concentration solutions to substrate and following the same extraction procedure as for samples. Run time for samples was 14 min with OTA being detected at about 11 min. The limit of quantification was 2.0 ng OTA g 21 CYA (ppb), while the limit of detection was 1.0 ng OTA g 21 CYA.
ELISA Ochratoxin A quantification was additionally performed with a competitive direct enzyme-linked immunosorbent assay (CD ELISA). Veratox ELISA quantitative test kit (Neogen Corp. Ltd, Lansing, USA) was applied to the samples, having a detection limit of 1 ppb (ng g 21 ) and range of quantification of 2-25 ppb. ELISA was strictly performed with accordance to the manufacturer's protocol and optical densities were determined at 650 nm absorbance with a spectrophotometer (Synergy HT, Biotek, USA). Quantification of OTA concentration was estimated with the aid of standard curves obtained from standard solutions provided within the test kit. When the determination of OTA by HPLC revealed concentrations greater than the upper limit of quantification by the ELISA method, serial dilutions of the extracts with the extraction solution took place in order to accomplish the ELISA method without extrapolation of its standard curves. Finally, non-ochratoxigenic isolates were omitted from screening with the ELISA method.
Among the potential OTA-producing fungi, only black aspergilli (Aspergillus section Nigri group) were isolated for further identification to species level. Initially, the 128 different black aspergilli species isolated from grapes were classified according to their morphological characteristics into three subgroups namely, Aspergillus niger aggregate, uniseriate species and A. carbonarius, and thereafter, a partial molecular identification took place.
Incidence of black aspergilli isolation from the sampled vineyards is presented in Figure 1. Regarding the presence of Aspergillus section Nigri, Macedonia presented very low numbers, namely 47% incidence (i.e., % of contaminated out of total analysed grapes) with 18% distribution in the isolated mycoflora (i.e., % of group in total recorded fungi), while between the other regions occurrence of black aspergilli was similar and always the dominant genera. The highest incidence and distribution were observed in Attica (99% incidence/55% distribution), followed by the Peloponnese (96%/34%) and Crete (76%/32%).
Regarding A. carbonarius, the highest distribution occurred for isolates from Crete, with 83% of black aspergilli identified, followed by the Peloponnese and Attica with 45% and 32%, respectively, while A. carbonarius isolation was most scarce in vineyards of Macedonia (15%). As regards uniseriate spp. a sole isolate was identified originating from Attica prefecture. According to the spore size that can discriminate between black uniseriate aspergilli, the isolate could be characterized as A. japonicus.

Molecular characterization of Aspergillus isolates by PCR-RFLP
In order to monitor the representation of black aspergilli (A. aculeatus, A. awamori, A. brasiliensis, A. carbonarius, A. foetidus, A. ibericus, A. japonicus, A. niger, A. tubingensis and A. uvarumn) reported to be isolated from grapes [20,25,34,37,38] we focused on a partial molecular characterization of the 128 isolates at the species level by use of PCR-RFLP of the 5.8S-ITS region. PCR mediated amplification of the ribosomal region was performed using the ITS1 and ITS4 primers resulting to amplicon molecular sizes of approximately 600 bp for all samples. RFLP analysis was done using the HhaI, HinfI and RsaI restriction endonucleases in order to differentiate among species. The restriction patterns of reference strains as presented in Figure 2 were used for comparison with the individual isolates restriction digestion patterns. HhaI and HinfI enzyme digestion can differentiate A. carbonarius from A. niger/A. tubingensis since they both generate 2 different restriction patterns designated here as A (for A. carbonarius) and B (for A. niger and A. tubingensis). RsaI digestion can distinguish further A. niger from A. tubingensis generating different restriction patterns for these two species, pattern A (common for A. niger and A. carbonarius) and pattern B (for A. tubingensis). A total of 128 black Aspergillus isolates previously characterized by morphological criteria were analyzed by PCR-RFLP and their restriction patterns were compared with those obtained from the reference strains. Digestion with HinfI resulted to 61 A and 66 B pattern profiles and 1 for a new profile designated as C and closely resembles profile B. HhaI digestion resulted to 60 A, 66 B pattern profiles and 2 for new pattern profiles designated as C and D. These new profiles for both enzymatic digestions can be seen for samples Aj-1 and Ai-1 in Figures 3 and 4. Restriction digestion with HhaI can also differentiate uniseriate (A. japonicus, A. aculeatus, A. uvarum) from all biseriate black aspergilli species [33,39]. Depending on the pattern which will result from the HinfI digestion, A. japonicus can be further distinguished from A. aculeatus [39]. According to restriction endonuclease site profile study (data not shown) of characterized black uniseriate aspergilli ITS sequences deposited in public sequence databases, HinfI digestion results in a double band visible pattern for either A. japonicus and A. uvarum and a three band visible pattern for A. aculeatus. The 66 strains which gave a B-B pattern combination after HinfI and HhaI digestions were subjected to digestion with the RsaI, resulting to 62 B and 4 A type restriction patterns. The strain with the C-C HinfI-HhaI, restriction digestion pattern when digested with RsaI, resulted to a new profile designated as C. This profile closely resembles RsaI profile A and can be seen in Figure 4. The strain with the A-D HinfI-HhaI, restriction digestion pattern, when digested with RsaI resulted to a pattern nearly indistinguishable from profile A (Figure 4). Though the molecular size of the largest DNA fragment generated (slightly above 500 bp) runs somehow faster that the corresponding band from profile A, the visible difference is not so obvious. This new profile was designated as D. According to the PCR-RFLP results for the 128 isolates, 60 (46.88%) were characterized as A. carbonarius, 62 (48.44%) as A. tubingensis, 4 (3.12%) as members of the A. niger aggregate, 1 (0.78%) -resembling the pattern C-C-C for HinfI, HhaI and RsaI digestions -could be characterized as A.
japonicus/A. uvarum, and 1 (0.78%) with an A-D-D HinfI, HhaI and RsaI digestion pattern could not be clearly classified to any of the previously referred species. A detailed classification according to both phenotypical and molecular data can be seen in Table 1.

Sequencing and Phylogeny study
Representative isolates (spp. Ac-28, At-10, Ac-36, Ai-1, An-2, Aj-1, At-46, Ac-57, An-4 and At-59) for each species as resulted from PCR-RFLP analysis were subjected to sequencing for results verification. A part from the resulted sequences alignment can be seen in Figure 5. The DNA sequence restriction pattern analysis was in accordance with these of reference strains for all isolates sequenced except isolate Ai-1. Isolate Ai-1 had a HinfI restriction site pattern common to A. carbonarius, however, the presence of an extra HhaI restriction site (nucleotides 143 to 146) generated a HhaI restriction site profile similar to this of A. japonicus. A presentation of the DNA fragments generated according to the HinfI, HhaI and RsaI restriction sites, in all sequenced isolates and representative reference isolates of black aspergilli species isolated from grapes can be seen in Table 2. The fragment sizes that are generated resemble the sizes of the fragments visualized after gel electrophoresis of the amplicon digests.
BLAST analysis of all sequences indicated that 3 clones had high similarity to A. carbonarius, 3 to A. tubingensis, 2 to A. niger/A. awamori, 1 to A. japonicus/A. aculeatus and 1 to A. ibericus. The results support the characterization of the 10 isolates deduced from the RFLP analysis and suggest that the isolate with the A-D HinfI-HhaI RFLP pattern can be classified as A. ibericus.
The phylogenetic relationship between the isolates of Aspergillus species is illustrated in an N-J method cluster analysis ( Figure 6). The ten sequences were aligned together with reference strains sequences retrieved from NCBI resulting to the formation of five clades. Isolate Aj-1 was aligned with uniseriate A. japonicus, A. uvarum and A. aculeatus forming together with A. aculeatinus, A. fijiensis and A. violaceofuscus a clade separated from all other biseriate Aspergilli sequences with a 100% bootstrap value. Two other major clades were formed. The first consisted by A. carbonarius/A. ibericus isolates (Ac-28, Ac-36, Ac-57 and Ai-1) together with A. carbonarius, A. ibericus and A. brasiliensis outgroup reference sequences, each one of the last three species forming a separate subgroup inside the clade. A. carbonarius sequences were aligned together and were separated from the A. ibericus sequences with a 98% bootstrap value. A. brasiliensis formed a separate branch in this clade with a 79% bootstrap value. The second major clade was formed by A. tubingensis/A. niger aggregate isolates (At-10, At-46, At-59 and An-2, An-4 respectively) and A. niger aggregate outgroup reference sequences. A. tubingensis sequences were clustered together with other black aspergilli (not reported to be found in grapes), forming a subgroup separated from all other A. niger aggregate (A. niger, A. awamori and A. foetidus) species of the clade, with a 86% bootstrap value. Two other minor clades were also formed, one by A.
saccharolyticus together with A. homomorphus and the other by A. ellipticus together with A. heteromorphus.

Ochratoxigenic potential of Aspergillus section Nigri species
A total of 128 isolates from grapes belonging to Aspergillus section Nigri, were tested in vitro for their potential to produce ochratoxin A. For the estimation of their ochratoxigenic potential a Competitive Direct ELISA procedure was also applied apart from the official HPLC method. Results showed that of the 60 A. carbonarius isolates 59 (98.33%) produced OTA from all geographic regions assayed. In contrast, among the A. niger aggregate, only 1  out of 62 isolates (1.61%), identified as A. tubingensis, had the potential to produce OTA. Finally, neither A. ibericus nor A. japonicus isolates were capable for OTA production.
Regarding the performance of the ELISA test kit quantification, results were linearly correlated with those of the HPLC method with an r 2 = 0.905 (Figure 7), but with an underestimation of the produced OTA at concentration levels above 5000 ppb. Results of both methods are presented in Table 1, together with the characterisation of the isolates after both phenotypic and molecular means.

Discussion
Incidence of isolation and distribution of Aspergillus section Nigri group among the fungal genera were in accordance with previous reports from Greece [20,51] and other Mediterranean countries [52,53,54]. A very high incidence was observed in our samples since 23 out of 33 vineyards sampled presented 100% presence for black aspergilli, although some studies have reported lower indices for vineyards from some European and South American countries [19,55,56,57].
Regarding the molecular characterisation of the isolated black aspergilli, a PCR-RFLP method was applied in order to differentiate the Aspergillus strains. RFLP analysis of the rDNA 5.8S ITS products was performed to identify different isolates of black Aspergillus species from grapes compared to reference strains. A total of 128 black Aspergilli isolate-ITS amplicons were digested using the HhaI, HinfI and RsaI restriction digestion endonucleases resulting in 4 different RFLP patterns for HhaI and RsaI respectively and 3 patterns for the HinfI enzyme digests. In previous studies [33,39,58,59,60]  According to the digestion pattern combinations resulted from the PCR-RFLP analysis of the 128 Aspergilli isolates and reference species, 60 were characterized as A. carbonarius, 62 as A. tubingensis and 4 as species belonging to the A. niger aggregate other than A. tubingensis. These results are in accordance with results presented by other authors, supporting the use of the HhaI, HinfI and RsaI restriction endonucleases for accurate characterization of A. carbonarius, A. tubingensis and, A. niger species [33,39,59,60]. Two other digestion pattern combinations have resulted using the HhaI and HinfI enzymes. One combination resumed a pattern presented by Spadaro et al. [39], suggesting that the isolate possibly belongs to the A. japonicus/A. uvarum species. The other digest pattern combination suggested a characterization of the isolate as A. carbonarius according to HinfI digestion; however, this could not be supported from the HhaI digestion profile which closely resembled this of A. japonicus.   Sequencing of representative isolate DNAs supported the RFLP data since the restriction site profile of each isolate generated DNA fragments with molecular sizes that corresponded to the bands found after electrophoresis of the amplicon digests. An extra HhaI site is present in the sequence of isolate Ai-1 generating a restriction digestion profile different from the rest that are observed. A nucleotide substitution, C instead of T that is present in nearly all other sequences generates this extra HhaI site. The results presented in Table 2 show that the DNA fragments generated according to the HinfI restriction sites are nearly equal in size and close to the DNA fragment sizes that generate pattern B for (A. tubingensis and A. niger). Though the pattern generated for A. japonicus is not identical to pattern B, it is very similar and cannot be distinguished clearly from it in a gel electrophoresis due to similar sizes and the techniques' limited resolution. A similar conclusion for unisereriate/biseriate species may be also deduced when evaluating by gel electrophoresis a profile generated after RsaI digestion.
Sequence alignment with reference strain sequences retrieved from NCBI, together with BLAST analysis verified the RFLP results and showed that 3 out of the ten individual isolates were assigned to A. carbonarius, 3 to A. tubingensis, 2 showed high homology to members of the A. niger aggregate (A. niger/A. awamori) and 1 showed high homology to A. japonicus/A. uvarum species. Isolate Ai-1 was assigned to the species A. ibericus after BLAST analysis, showing 100% homology when aligned with A. ibericus AY656624.1, AY656623.1 and AY656622.1 strains. A. ibericus was first reported by Serra et al. [61] as a new Aspergillus species isolated from grapes and to our knowledge this is the first report that this species is isolated in Greece. Furthermore, as shown from   a comparison of the A. ibericus 5.8S-ITS DNA sequences available at NCBI (data not shown), the fact that an extra HhaI site is present in all of these sequences makes it useful for characterization of A. ibericus by PCR-RFLP using this restriction endonuclease. The phylogenetic analysis performed using reference species' sequences as outgroups supports the sequence data and is in accordance with data presented by other authors using the ITS region for phylogenetic analysis of black aspergilli [37].
According to various authors [20,34,37,38,58,62,63,] black aspergilli that have been isolated from grapes fall into the biseriate A. carbonarius, A. tubingensis, A. niger, A. awamori, A. foetidus, A. brasiliensis and A. ibericus, and the uniseriate A. japonicus, A. uvarum and A. aculeatus species. Characterization of these isolates by molecular means has been based on different approximations using RFLP [25,31,33,39,64], AFLP [30,32,36,65], or multilocus analysis [62,66]. RFLP and AFLP have been used successfully for characterization of the main ochratoxigenic aspergilli isolated from grapes [25,30,31,32,33,36,39,64,65]. Multilocus analysis based on calmodulin, beta-tubulin and ITS sequences has proved a useful tool for extended characterization and comparative study of isolates from a great variety of sources such as grapes, coffee, soil, surfaces, or unknown origin [37,62,66]. In this work we have followed a similar approximation using ITS PCR-RFLP in order to characterize black aspergilli isolates from grapes. We have been able to characterize the isolates belonging to the A. carbonarius, A. tubingensis and A. ibericus species, since the RFLP patterns (Table 2) can clearly differentiate these species from all other black aspergilli species isolated from grapes.
The Aj-1 isolate could not be characterized only by molecular means since there is no difference between the RFLP patterns and ITS region DNA sequence for either species. According to the morphological observations the Aj-1 isolate was characterized as A. japonicus since its conidial size (4.3 mm) falls in the range of A. japonicus (4-5 mm) and not in the range of A. uvarum (3-4 mm) [34]. PCR-RFLP when used in combination with other means such as use of morphological criteria is a reliable approximation for black aspergilli characterization, where, for example, differentiation of aspergilli as biseriate/uniseriate can initially short the strains to two major groups. Isolates An-1 to An-4 have been characterized as A. niger aggregate members due to the limitation of the ITS PCR-RFLP to differentiate between A. niger, A. awamori and A. foetidus species, thus resulting to a partial molecular characterization of these four isolates.
With regard to the field study, Greece is a country with extended coast, although, areas of grape planting are characterized by different climatic characteristics. Vineyards localisation can vary from very low altitude, near seaside, either arid or wet, to very high altitudes and wet climate regions. Referring to the present study, influence of the geographic localization of the vineyards on the incidence of black aspergilli in grapes was significant. The vineyards selected for this study are representative of these different climatic profiles existing in the country. Most studies that correlated the geo-climatic conditions with the presence and distribution of black aspergilli reported a trend of higher isolation incidence in regions with higher mean temperatures, but more wet areas [19,53,66,67]. These reports are partly in accordance with the present study, since the geographical localisation of sampled vineyards influenced black aspergilli presence, but with the higher incidences found in arid areas. In accordance with our results are the findings of Serra et al. [54,56] for Spain where higher incidences for black aspergilli were reported from hot and dry regions. In addition, most of the ochratoxigenic isolates came from regions with arid climate, Crete and Attica prefecture, but this could be also correlated with the low altitude of these vineyards in contrast with the Peloponnese and Macedonia where altitudes of grape cultivation are higher. Moreover, the incidence of black aspergilli and the A. carbonarius isolation were much lower in Macedonia where temperatures are lower and the climate is wet compared with the other regions. In contrast, Chiotta et al. [65] have reported higher incidences of A. carbonarius from vineyards of higher altitudes. In addition, Perrone et al. [20] and Visconti et al. [24] denoted the strong correlation of relative humidity and rainfall with A. carbonarius presence and OTA accumulation in grapes. Finally, many works denote that, apart from the climate of a wider region, very important role have also the different practices used in grape cultivation and the specific microclimate of the vineyard [68,69,70].

CYA)
Prefecture n (OTA producers)  A. niger aggregate and A. carbonarius were equally distributed in the present work, while only one uniseriate species was isolated and identified as A. japonicus. Similar distributions of A. carbonarius within the A. section Nigri group has been also reported by Bau et al. [69] for grapes from the Mediterranean Spanish coast and Tjamos et al. [51] for Rhodes island in Greece. These results are in accordance with our findings, since Crete, which has an insular Mediterranean climate, similar to those of Rhode and the Mediterranean Spanish coast, presented the higher incidence of A. carbonarius among black aspergilli. In contrast, El Khoury et al. [71] for Lebanon and Chiotta et al. [65,66] for Argentina have reported relatively lower participation of A. carbonarius within black aspergilli, while Battilani et al. [19] reported very high occurrence (20%) of uniseriate species in Italian vineyards during grape ripening.
Several studies from countries all over the world during the last years have identified A. section Nigri group as the responsible fungi for OTA contamination of grapes and derived products. Specifically, A. carbonarius is considered as the main OTA contamination source, and to a lesser extend species of A. niger aggregate, mainly A. niger and A. tubingensis. This conclusion derives from the higher ochratoxigenic ability of A. carbonarius in contrast to A. niger aggregate isolates that, even when they are more abundant within the A. section Nigri group, they are either totally incapable of toxin production or their toxigenic ability is scarce and weak [19,53,72]. In our study, all A. carbonarius isolates, with one exception, were capable for OTA production with most of them having a high ochratoxigenic potential, while from A. niger aggregate only one isolate identified as Aspergillus tubingensis was capable to produce OTA in vitro (Table 3). Similar results have been reported for other Mediterranean countries [19,54,56,69,73], although there are reports of higher ochratoxigenic incidences among A. niger aggregate and uniseriate spp. [39]. Ranges of ochratoxigenic potential for A. carbonarius isolates varied from 0 to 34 ppm (mg OTA g 21 CYA), with most of them producing between 1 and 1000 ppb (Table 3). Our results are also in agreement with studies contacted in Greece [51,68], although Greek isolates with higher ochratoxigenic potential have been reported [31]. Particularly, Tjamos et al. [51,68] in their studies have also shown that most A. carbonarius isolates were strong OTA producers with capabilities greater than 25 ppb, while only 6% of the A. niger aggregate presented high OTA potential.

Conclusions
From our results there is a clear correlation of A. section Nigri group with the arid climate of Crete and Attica, whereas for colder and wet areas its presence is scarcer. Moreover, higher incidence of ochratoxigenic isolates and OTA production was recorded for the A. carbonarius isolates, confirming in this way that A. carbonarius is the main species responsible for OTA contamination in Greek grapes. Regarding PCR-RFLP our data indicate that A. ibericus can be differentiated from other aspergilli using the HhaI restriction endonuclease.