Molecular characterization and sensitivity to demethylation inhibitor fungicides of Aspergillus fumigatus from orange-based compost

Aspergillus fumigatus, the causal agent of human aspergilloses, is known to be non-pathogenic in plants. It is present as saprophyte in different types of organic matter and develops rapidly during the high-temperature phase of the composting process. Aspergilloses are treated with demethylation inhibitor (DMI) fungicides and resistant isolates have been recently reported. The present study aims to estimate the abundance, genetic diversity and DMI sensitivity of A. fumigatus during the composting process of orange fruits. Composting of orange fruits resulted in a 100-fold increase in A. fumigatus frequency already after 1 week, demonstrating that the degradation of orange fruits favoured the growth of A. fumigatus in compost. Most of A. fumigatus isolates belonged to mating type 2, including those initially isolated from the orange peel, whereas mating type 1 evolved towards the end of the composting process. None of the A. fumigatus isolates expressed simultaneously both mating types. The 52 investigated isolates exhibited moderate SSR polymorphisms by formation of one major (47 isolates) and one minor cluster (5 isolates). The latter included mating type 1 isolates from the last sampling and the DMI-resistant reference strains. Only few isolates showed cyp51A polymorphisms but were sensitive to DMIs as all the other isolates. None of the A. fumigatus isolates owned any of the mutations associated with DMI resistance. This study documents a high reproduction rate of A. fumigatus during the composting process of orange fruits, requesting specific safety precautions in compost handling. Furthermore, azole residue concentrations in orange-based compost were not sufficient to select A. fumigatus resistant genotypes.


Introduction
The production of compost recently gained much importance in the EU as a waste management process. Composting reduces the volume of waste going to landfills, and the related CH 4 and CO 2 emission due to organic material degradation [1]. Composting process allows to upcycle a variety of organic materials and biowastes originating from different sources. The process begins with a thermophilic phase, when temperature reaches 55-65˚C, followed by a maturing and mesophilic stage when the temperature declines and the material stabilizes. Reaching a temperature of 55˚C for at least 3 days is an important step to guarantee sanitization and microbial safety of compost products [2]. It is specifically recommended by compost quality regulations applied in several EU countries.
The opportunistic fungal pathogen Aspergillus fumigatus (Fresen., syn. Neosartorya fumigata, O'Gorman, Fuller and Dyer), causal agent of the "aspergilloses" in immunocompromised human hosts, is known to be present in different types of compost [3][4][5][6]. It can survive successfully under a wide range of environmental conditions, particularly due to its capacity to resist high temperatures (thermo-tolerant).Consequently, it is prevalent during the high-temperature phase of the composting process and it can be present in compost at concentrations of up to 10 6 −10 7 cfu/g dw [5,7,8]. The spores are released to the air during composting activities such as heat turning, reaching concentrations of 10 4 −10 7 /m 3 air [9,10]. Although it can be present in relatively high concentrations, the current hygienic EU regulations for compost do not require an estimation of its presence but rather seek for the absence or limited presence of other human pathogens such as Salmonella spp. and Escherichia coli [1].
Demethylation inhibitor fungicides (DMIs, chemically named also as 'azoles') are the principal antifungal compounds used in the treatment (prophylactic and curative) of human and animal diseases caused by A. fumigatus [11]. DMIs inhibit the enzyme lanosterol 14-α demethylase (encoded by cyp51 gene, A and B paralogues), involved in ergosterol biosynthetic pathway of fungi [12,13]. Chemically related molecules have also been used for fruit pre-and post-harvest treatments and application of field crops in agriculture for more than 30 years [14,15]. Resistance to medical triazoles (e. g. itraconazole) is well-known since the 1990s and associated with several mutations in the cyp51A gene of A. fumigatus occurring in patients having been treated with these antifungal compounds [16]. Recently, triazole resistant strains have been isolated also from patients not previously treated with triazoles, suggesting that they have picked up air spores originating from any kind of DMI treated material in agriculture and general environment [3,17]. Since compost is one of the most important environmental sources for A. fumigatus [14], DMI resistant strains might have their origin from this substrate contributing to human health risks [3]. Recently, many azole resistant strains of A. fumigatus containing a new resistance mutation with clinical relevance were found in compost produced from conventional flower bulbs leftovers [6]. However, it is still not clear whether these mutants evolved during the composting process (e.g. by DMI residues) or were selected by earlier DMI treatment of plant material which was then used to produce the compost pile. Indeed, we previously investigated the presence of A. fumigatus isolates in commercial composts produced from different wastes, with different composting methods in different locations (Italy, Spain, Hungary, The Netherlands, Germany and United Kingdom): A. fumigatus was present in all composts but not a single isolate shown mutations for DMI resistance [4,5].
The aim of this study was to estimate the presence and abundance of A. fumigatus during the composting process of orange fruits, commonly treated with post-harvest fungicides like imazalil (a DMI fungicide), and to evaluate whether this biowaste may be considered as a "hot spot" for DMI resistance evolution. Furthermore, the genetic diversity of isolates, their sensitivity to DMIs and the molecular characterization of the coding part of the cyp51A gene were assessed, in order to identify possible mutations occurring during the composting process. Our hypothesis was that the composting process of orange fruits would allow A. fumigatus to multiply and reproduce, but that the low azole concentrations present in the orange peel would not select for azole-resistant A. fumigatus genotypes.

Collection of oranges samples, identification and quantification of A. fumigatus isolates
Oranges samples (around 300 orange fruits) have been collected in Ivrea among leftovers from the local carnival. Oranges were specifically produced for the carnival in the area of Corigliano Calabro (Southern Italy, GPS coordinates: 39˚40'21.7"N 16˚25'42.0"E), and not used for food purposes. The oranges were routinely treated by imazalil in order to prevent the development of post-harvest fungal diseases, with a maximum residue level <5 mg per kg of orange fruits according to EU regulation (http://ec.europa.eu/food/plant/pesticides/eu-pesticides-database/ public/?event=homepage&language=EN). Twenty percent of the oranges have been examined for the presence of A. fumigatus, the rest was delivered to an industrial composting plant located in Piedmont. Isolation and identification of A. fumigatus was carried out following the protocol by Franceschini et al. [4]. One g of three subsamples (orange peel) was added to 9 ml Ringer solution and shaken for 1 hour at 100 rpm. One hundred μl of suspension and serial dilutions of it were plated on 20 ml of potato dextrose agar (PDA, Merck1, Darmstadt, Germany) containing streptomycin (50 mg/l, Applichem, Darmstadt, Germany), in triplicates. Plates were incubated at 42˚C for 3-5 days.
Typical powdery green-grey colonies of two isolates (AR4.9 and AR6.1) from two orange fruits were purified, and identified by macro and micro-morphological characteristics. The plates were incubated at 50˚C with the aim to distinguish A. fumigatus from other Aspergillus species. Colonies capable to grow at 50˚C were counted on all plates, multiplied by the dilution factor and converted to one gram of dry weight (dw) of the orange peel. Monoconidial cultures obtained from each isolate were stored in 30% glycerol solution at -80˚C.

Preparation of orange-based compost
The oranges were composted using a bench-scale system consisting of 3.5-L PVC plastic units incubated at 50˚C for 1 month starting from March 6 th 2017. To do so, four kg of oranges were crushed and added to each unit embedded in a plastic bag to avoid external contaminations. The composting process started with a first phase of active degradation lasting nearly 7 weeks, followed by another 4 weeks of maturation phase at a pH > 7. Compost samples were collected during 10 sampling dates (86 days in total), at a 7-14 days interval, and the composting material was well mixed every time (Table 1).

Identification and quantification of A. fumigatus isolates
Fifty A. fumigatus isolates (5 isolates per each sampling date) were collected during the composting process and identified as described above (Table 1). For isolation of A. fumigatus from compost, we used a selection medium without azoles in order to pick up isolates with reduced sensitivity to DMIs as well as fully sensitive isolates for further in vitro sensitivity assays.
Fifty isolates from compost together with the two isolates from intact orange fruits were used subsequently for molecular identification using internal transcribed spacer (ITS) region, microsatellite analysis, sensitivity assays to DMIs, and molecular characterization of the cyp51A gene. Fungal DNA was extracted by the EZNA1 Fungal DNA extraction kit (Omega Bio-Tek, Darmstadt, Germany) following the manufacturer's instructions. The amplification of the ITS region was performed as described by White et al. [18], and the sequences of the ITS amplicons were used for molecular identification.

Mating types identification
Mating type of A. fumigatus isolates was determined by using multiplex PCR according to Paoletti et al. [19]. Multiplex PCR assay included MAT-1 specific primer (AFM1,

Genetic diversity assessment of A. fumigatus isolates
Six microsatellite regions (STRAf 3A, 3B, 3C, and STRAf 4A, 4B, 4C) of 52 A. fumigatus isolates were amplified using the set of primers described by Valk et al. [20]. Four reference environmental isolates (provided by B. Fraaije, Rothamsted Research, Harpenden, UK) were also included in the analysis: wild-type isolate (WT), and resistant isolates (TR 34 +L98H from UK, TR 46 +Y121F+T289A from UK and TR 34 +L98H from NL). The PCR mixture (20 μl) contained 1 ng of genomic DNA, 1× PCR buffer, 1 μM of each primer, 0.2 mM dNTPs, 2.0 mM of MgCl 2 , and 1 U Taq DNA polymerase (Qiagen). PCR conditions were as follows: initial denaturation at 95˚C for 10 min, followed by 30 cycles of denaturation at 95˚C for 30 s, annealing at 60˚C for 30 s and extension at 72˚C for 1 min, and final extension at 72˚C for 10 min. STRAf 3B and STRAf 4B microsatellite loci were amplified using an annealing temperature of 65˚C. Amplified DNA fragments were separated on 3% MetaPhor1 agarose gel (Lonza, Rockland, USA), and then analyzed by PyElph 1.4 software [21]. The calculation of genetic distance between isolates was performed by GenAlEx 6.502 software [22,23]. Cluster analyses of polymorphism detected by six SSR markers were performed using the Unweighted Pair Group Method with Arithmetic Averages (UPGMA) of the MEGA 7 software [24].

In vitro sensitivity testing of A. fumigatus isolates to DMIs
Fifty-two A. fumigatus isolates were tested for in vitro sensitivity to the DMIs by using one post-harvest and veterinary medical DMI (imazalil, syn. enilconazole) and two human medical DMIs (posaconazole and voriconazole). EUCAST protocol [25] was applied with slight modifications (incubation temperature at 37˚C, and visual reading of results using turbidity scale) following the FRAC standard methods by in vitro assays for plant pathogens on microtiter plates (http://www.frac.info/monitoring-methods). Four environmental isolates were additionally tested as the reference isolates (WT isolate and 3 resistant isolates as indicated above). Imazalil (PESTANAL1 analytical standard, Sigma-Aldrich), voriconazole, and posaconazole (VETRANAL™ analytical standard, Sigma-Aldrich) were tested using 5-fold serial dilutions (from 50 to 0.08 mg/l), in two replicates. RPMI 1640 medium (with L-glutamine, Sigma-Aldrich) containing 2% glucose, 3-(N-morpholino) propanesulfonic acid (MOPS; final concentration of 0.165 mol/l, pH 7.0), and each specific fungicide was loaded (100 μl/well) into flat-bottom Nunc™ 96-well microplate (Thermo Fisher Scientific, Wilmington, USA). One hundred μl of A. fumigatus spore suspension (2-5 × 10 5 conidia/ml) was then added to each well. Appropriate controls (without fungicide or without A. fumigatus) were also included. The microplates were incubated for 48 hours at 37˚C. Mycelial growth of all isolates was visually assessed [25] by grading turbidity scale (0-5) with 0 referring to optically clear and 5 indicating no reduction in turbidity compared with that in the control (fungicide-free) well.
The percent growth inhibition (GI) was calculated as: % GI = (Gc-Gf / Gc) x 100, where 'Gf' indicates the growth percentage at each fungicide concentration, and 'Gc' refers to the growth control. EC 50 values (or concentrations causing 50% growth inhibition) were calculated using a 'log/logit dose response' parameter of the GraphPadPrism1 software (version 7.02; La Jolla, CA, USA). The log fungicide concentration vs. normalized response-variable (GI) procedure using a logistic regression was calculated as: Y = Bottom + (Top-Bottom) / {1 +10 [(LogEC 50 -X) × HillSlope]}, with Y indicating the response, X referring to the fungicide concentration, Top and Bottom indicating the plateaus in the Y axis units, and Hillslope refering to the steepness of the curve [26]. EC 50 values were used rather than MIC values due to their very precise calculation of the dose-response relationship (see FRAC guidelines).
The sensitivity distribution or frequency distribution of isolates at certain EC 50 level was determined for imazalil, posaconazole and voriconazole using the 'box-and-whiskers' plot method. This method provides the variability, shape, and characteristic distribution values through the maximum and minimum values ('whiskers'), the interquartile range (the differences between the various interquartiles), box (50% of population), the median (central line) value, and the average value [27].

Molecular characterization and sequence analyses of the cyp51A gene
The full cyp51A gene of 52 A. fumigatus isolates was amplified according to Snelders et al. [28] using the primers P450-A1 (5'-ATGGTGCCGATGCTATGG-3') and P450-A2 (5'-CTGTC TCACTTGGATGTG-3'). The PCR cycling conditions consisted of an initial denaturation of 5 min at 95˚C, followed by 40 cycles at 94˚C for 30 s, 58˚C for 45 s, and 72˚C for 2 min, and a final extension at 72˚C for 7 min.
The ITS and cyp51A PCR products were sequenced at BMR Genomics (Padua, Italy). The GenBank accession numbers of the ITS and cyp51A sequences are reported in Table 1. BLAST searches of obtained sequences were performed against the GenBank databases at NCBI. Multiple sequence alignment was performed by Vector NTI Advance 11 software (InforMax, North Bethesda, Maryland, USA) using the Clustal W algorithm [29]. The cyp51A sequence of the WT isolate (accession no. AF338659), sensitive compost isolates [4,5] and resistant clinical and environmental isolates (ITZ.86, 11_0087A, 14, and 98) [28,[30][31][32] were included in the analyses. Phylogenic analyses were carried out using MEGA 7 software, creating the neighbour-joining (NJ) trees with 1000 bootstrap replications.

Morphological and molecular identification of A. fumigatus on orange peel
Out of 60 orange fruits, two A. fumigatus isolates (AR4.9 and AR6.1) were detected on the peel of two different asymptomatic orange fruits, and identified morphologically (frequency of 3.3% on fruits) corresponding to a concentration of 0.01 to 0.04 × 10 3 cfu A. fumigatus per g of dw of orange peel. The two isolates AR4.9 and AR6.1 were confirmed as A. fumigatus by molecular characterization on the basis of ITS sequencing (GeneBank Accession Nrs. MG976901 and MG976902).

Identification and quantification of A. fumigatus isolates in orange-based compost
In total, 50 orange-based compost isolates of A. fumigatus from different composting phases were collected and identified on the basis of macro and micro morphological observations. During each sampling date, one compost sample was collected and divided into three subsamples. Five randomly selected colonies per each sample (including all three subsamples) were confirmed as A. fumigatus by ability to grow at 50˚C and sequencing ITS amplicons ( Table 1). The abundance of isolates was measured and expressed as cfu/g dw of compost ( Fig 1A). The values at the beginning of the composting process were at 8.8 × 10 3 cfu/g, increased more or less steadily and reached highest concentrations on a plateau level at 430.7 to 605.7 × 10 3 cfu/g at the last three dates of sampling (Fig 1A).
In addition, the pH of compost samples was measured; it was 3.6 at the beginning and increased steadily during the composting process reaching a value of 7.8 at the end (Fig 1B). The fungus was able to grow under both acidic and slightly basic conditions, with the highest abundance at neutral to slightly basic conditions during the second half of composting when the composting process was almost accomplished.

Mating types identification
Out of 52 A. fumigatus isolates, six expressed mating type 1 (12%) and 46 mating type 2 (88%) (S1 Fig and Table 1). No isolate showed amplification of both mating-type bands. Interestingly, all five isolates originating from the last (10 th ) sampling were of mating type 1, the other mating type 1 isolate came from the seventh sampling period.

Genetic diversity of A. fumigatus isolates
Based on six SSR markers, the genetic diversity within the population of 50 A. fumigatus isolates from orange-based compost and 2 isolates from orange fruits was moderate, with two main clusters (Fig 2). The first cluster contained most (47) of the isolates and all isolates that were sensitive to DMIs. The second cluster included 3 reference DMI resistant isolates and the orange-based compost isolate 502.3_24.5 that showed reduced sensitivity to voriconazole and posaconazole (Table 2). Three resistant reference isolates were in a separate sub-cluster, while the orange-based compost isolate 502.3_24.5 was distant from the other isolates. Interestingly, all A. fumigatus isolates from the last sampling date and the reference wild-type isolate were also included in the second cluster. Moreover, the sub-clusters in both clusters contained mainly isolates from the same or close sampling periods.
Fifty-two A. fumigatus isolates were fully sensitive to the tested DMI fungicides, with exception of one isolate (502.3_24.5) in combination with two mecidal triazoles (voriconazole and posaconazole). This isolate was the only one among five isolates from the same (and the last) sampling date with an increased EC 50 value. Two isolates (502.2_24.4, and 502.2_24.5) that showed polymorphisms at aa positions 46, 120, 172 and 427 were also fully sensitive to the three tested fungicides.
The 'box-and-whiskers' plots including one wild-type and three resistant reference isolates showed important sensitivity variations in the 50% box of population for posaconazole, and high maximum whiskers values for voriconazole and posaconazole. The sensitivity of the orange-based compost isolate 502.3_24.5 was clearly outside the 50% box for voriconazole and posaconazole. The highest whiskers value was observed for posaconazole. The most uniform sensitivity distribution taking in consideration all calculated parameters was found for imazalil (Fig 3).

Molecular characterization of the cyp51A coding gene sequence
Out of 14 A. fumigatus isolates from orange-based compost, two isolates (502.2_24.4 and 502.2_24.5) belonging to two different composting periods (24 April and 24 May) showed 4 shared cyp51A polymorphisms: F46Y, V120G, M172V, and E427K (Fig 4). The isolates 502.5_27.3 and 502.2_24.4 shared the mutation S52T, and the isolates 502.3_24.4 and 502.2_24.5 the mutation F495L. Single aa substitutions were also found in isolate 502.2_24.4 at positions 115 (F to V), 142 (S to P) and 154 (E to D), and in the isolate 502.2_24.5 at position 496 (S to T). On the other hand, 10 other isolates did not show any polymorphisms, including the isolate 502.3_24.5 that showed reduced sensitivity to voriconazole and posaconazole ( Fig  4). Also, the two A. fumigatus isolates from orange fruits (AR4.9 and AR6.1) had no aa substitutions. Mutations known to be responsible in DMI resistance (L98H, Y121F, and T289A) were present in the resistant reference isolates but not in any of the sequenced A. fumigatus isolates from compost, and orange fruits.
Moreover, all isolates from this and previous studies [4,5] were aligned in the coding region of the cyp51A gene (515 aa). Phylogenetic analyses grouped the isolates in two clusters (S2 Fig). The main cluster contained all orange-based compost isolates, WT and resistant reference isolates, while the second cluster included the compost isolates with a higher number of cyp51A polymorphisms.

Discussion
A. fumigatus is not a plant pathogen and does not invade living plant cells. However, it can grow easily as saprophyte in decaying organic matter. Presence of A. fumigatus has also been associated with the spoilage and loss of citrus fruits during storage [33]. Several Aspergillus spp. have been reported as epiphytes on the peel of citrus fruits [34,35]. During the composting process of orange fruits A. fumigatus population was found to increase a 100-fold already after 1 week, confirming the degradation power for organic matter at high temperatures of this thermo-tolerant fungus [7]. The maturing phase of composting in which the temperatures were still high and pH changed from acidic to slightly basic, caused a further increase of fungal growth, demonstrating that degradation of orange fruits favoured the presence and abundance of A. fumigatus in compost. Similarly, abundance of A. fumigatus increased tremendously in corn silage already one week after silo opening [36]. Therefore, it can be expected that in all organic substrates undergoing a thermophilic phase during degradation (eliminating a big portion of the microbial populations) A. fumigatus may be present in rather high quantities. This represents a quite high health risk especially for immunocompromised people and those who are sensitive to allergic reactions. In addition, the fungus may contaminate the organic substrates through the production of mycotoxic compounds such as fumigacin [14].
N. fumigata, the teleomorph (sexual) stage of A. fumigatus, is probably rare under natural conditions and only recently discovered [37]. Genes associated with the expression of both mating types are present in the genome of A. fumigatus. Mating type 1 is considered to be associated to higher invasiveness of A. fumigatus in patients, whereas mating type 2 is more frequent in environmental populations [38]. In our study, none of the isolates expressed simultaneously both mating types, most of them belonged to mating type 2, including those isolated from orange peels, whereas mating type 1 evolved towards the end of the composting process. This pattern demonstrates that also mating type 1 isolates can develop in environmental populations even if there was a predominance of mating type 2 at the beginning of the process.
The analysis of the SSR genetic diversity showed that the five isolates from the final composting process belonged to mating type 1 also clustered together with three DMI resistant reference isolates in cluster II. Moreover, many of the cluster I isolates, including one other mating type 1 isolate harbored several polymorphisms in cyp51A gene but were sensitive to the three tested fungicides, suggesting that, in the presence of low azoles concentrations, composting of orange fruits would not select for azole-resistant A. fumigatus genotypes nor for the well known mutations coding for triazole resistance (L98H, Y121F, T289A).
A. fumigatus isolates from orange-based compost (this study) exhibited moderate SSR polymorphisms supporting microsatellite genotypic results of other environmental populations [39] and commercial compost [5]. Furthermore, it was not possible to separate DMI-sensitive and DMI-resistant strains. Although three resistant reference strains grouped together in the same sub-cluster, they were together with a wild-type reference isolate. Thus, SSR genotyping is not related to fungicide sensitivity as was already observed in previous studies [5,30,39]. However, clustering might be related to the geographic origin as shown earlier by Santoro et al. [5] and to a specific composting phase, since all isolates from mature compost were grouped together in one cluster, while isolates from the earlier composting phases were in another cluster. Our data also showed that SSR markers were polymorphic even within the Genetic diversity and DMI sensitivity of Aspergillus fumigatus from orange-based compost Genetic diversity and DMI sensitivity of Aspergillus fumigatus from orange-based compost same sample suggesting a high genetic drift or frequent mutations happening from one generation to the next. Also, A. fumigatus population may not simply grow clonally, but may undergo continuous sexual or parasexual recombination as was already reported earlier [40]. Few A. fumigatus isolates in this study showed cyp51A polymorphisms but were sensitive to DMIs as was reported already in environmental isolates: F46Y from soil and indoor/outdoor hospital environment [3,28]; S52T from seeds and compost [3,5]; F115V, V120G, S142P and E145Q from compost [5]; M172V from soil and indoor/outdoor hospital environment [3,28]; E427K from soil, indoor/outdoor hospital environment, and compost [3,4,28]; and F495L from indoor/outdoor hospital environment [28]. These polymorphisms were reported also in sensitive clinical isolates (reviewed by Stensvold et al. [41]). None of the orange-based compost isolate in this study owned any of the cyp51A mutations related to DMI resistance, similarly to our previous findings on compost and biochar isolates with different geographical origins (Italy, Spain, Hungary, The Netherlands, Germany and United Kingdom) [4,5]. The cyp51A amino acid changes found in this study more likely present particular genotypes related to Genetic diversity and DMI sensitivity of Aspergillus fumigatus from orange-based compost specific developmental stage (soil, compost) rather than specific locations. Taking all isolates from this and earlier studies [4,5] together, there was no cyp51A clustering pattern according to geographic origin (countries) (see S2 Fig). On the other hand, the isolate 502.3_24.5 expressed reduced sensitivity to both medical triazoles, voriconazole and posaconazole (although no cyp51A mutation was detected), but was fully sensitive to imazalil. This may suggest that this isolate either has been selected preferentially by the medical triazoles (preferential selection is known within DMIs in certain plant pathogens, [42]), or could express resistance mechanisms that are specific for the two medical triazoles (e.g. ABC pumps). It is still under debate whether DMI resistance in A. fumigatus has primarily medical origin (due to medical/veterinary DMI treatments of aspergilloses) or environmental and/or agricultural origin (due to DMI applications against plant pathogens and in material protection). Our findings so far suggested that A. fumigatus of environmental origin is not a source for cyp51A mutations associated with DMI resistance. The same findings were observed for environmental A. fumigatus isolates from compost and biochar in our previous studies [4,5] and in other similar investigations [43][44][45]. DMI resistance observed in A. fumigatus isolates of environmental and clinical origin in other studies [6,[30][31][32]39,[46][47][48] might be associated to much higher azole concentrations in the substrate (and subsequently higher selection pressure) or to certain geographic locations.
Composting obviously presents a very favourable niche for A. fumigatus to grow and multiply, however the process itself might be disadvantageous for resistance evolution probably due to degradation or adsorption of fungicides in the organic substrate [49]. Since imazalil is not as persistent in soil environment (dissipation half-life (DT 50 ): 30-170 days, [50]) and compost at low pH (DT 50 : 55-120 days, [51,52]) as certain agricultural triazoles, it can be assumed that its residue quantities in the orange-based compost are not sufficient for a possible selection of resistant individuals. Consequently, composting orange fruits would allow A. fumigatus to reproduce, but not to select azole-resistant genotypes. Furthermore, our study demonstrates a high genetic diversity in A. fumigatus isolates even when deriving from the same environmental sample, suggesting frequent sexual or para-sexual recombination events within the population. However, it was not possible to genetically distinguish sensitive environmental from resistant reference isolates because they were within the same clusters. Finally, more emphasize should be given to the presence of A. fumigatus in compost in order to achieve better safety precautions for people in close contact with compost (e.g. compost operators).

Conclusions
In the present study, the abundance, genetic diversity and DMI sensitivity of A. fumigatus was evaluated during the composting process of orange fruits. Although composting presents a very suitable niche for A. fumigatus grow, it may be assumed that imazalil residue concentrations are not sufficient to select resistant genotypes during orange-based composting.