Epidemiological Studies of Pan-Azole Resistant Aspergillus fumigatus Populations Sampled during Tulip Cultivation Show Clonal Expansion with Acquisition of Multi-Fungicide Resistance as Potential Driver

Pan-azole resistant isolates are found in clinical and environmental Aspergillus fumigatus (Af) populations. Azole resistance can evolve in both settings, with Af directly targeted by antifungals in patients and, in the environment, Af unintendedly exposed to fungicides used for material preservation and plant disease control. Resistance to non-azole fungicides, including methyl benzimidazole carbamates (MBCs), quinone outside inhibitors (QoIs) and succinate dehydrogenase inhibitors (SDHIs), has recently been reported. These fungicide groups are not used in medicine but can play an important role in the further spread of pan-azole resistant genotypes. We investigated the multi-fungicide resistance status and the genetic diversity of Af populations sampled from tulip field soils, tulip peel waste and flower compost heaps using fungicide sensitivity testing and a range of genotyping tools, including STRAf typing and sequencing of fungicide resistant alleles. Two major clones were present in the tulip bulb population. Comparisons with clinical isolates and literature data revealed that several common clonal lineages of TR34/L98H and TR46/Y121F/T289A strains that have expanded successfully in the environment have also acquired resistance to MBC, QoI and/or SDHI fungicides. Strains carrying multiple fungicide resistant alleles have a competitive advantage in environments where residues of multiple fungicides belonging to different modes of action are present.


Introduction
Airborne Aspergillus fumigatus (Af ) spores can, after inhalation, cause disease in animals and humans ranging from allergic conditions and chronic lung infection to acute invasive aspergillosis (IA). Patients with a weakened immune system are most at risk for IA, and successful therapy depends on early diagnosis and the effective use of antifungals. Due to their efficacy and low toxicity, orally administered azoles have been the drugs of choice, followed by intravenous applications of amphotericin B and echinocandins. However, azole resistance has evolved in Af and is becoming more common in both the clinical setting and the wider environment since the late 1990s [1][2][3]. Further spread of resistance will have a significant clinical impact as higher mortality rates have already been recorded for patients with voriconazole-resistant IA in comparison with voriconazole-susceptible infection [4]. 1 Isolates were obtained from colonies growing on Sabouraud Dextrose agar plates without fungicides with exception of the compost heap isolates that were picked from plates amended with 5 µg/mL tebuconazole. 2 Azole resistant isolates have at least two raised MIC values for either voriconazole (>1.0 µg/mL), imazalil (>2.0 µg/mL) and/or tebuconazole (>4.0 µg/mL).
Materials from tulip peel waste heaps and compost sites were sampled and sent in by commercial tulip growers during August 2018. For both soil, bulb peel waste heap and compost samples, 2 g aliquots of material were added to 8 mL of phosphate buffered saline amended with 0.1% (v/v) Tween 20 (PBST) and Af strains isolated as described in our previous study [17]. Tulip and daffodil bulbs were purchased directly from garden centers or ordered on-line from nurseries in the autumn during 2016-2018. The outer skin of ten bulbs per sample was peeled off, pooled, and further processed using 25 mL of PBST. To grow and isolate Af colonies aliquots of the buffer extracts were plated out on untreated, or azole fungicide (5 µg/mL tebuconazole) amended Sabouraud dextrose (SD) agar (Oxoid Ltd., Basingstoke, UK) containing penicillin (100 U/mL) and streptomycin (100 µg/mL) and incubated for 2 days at 48 • C.

Fungicide Sensitivity Testing
After subculturing single colonies in tissue culture flasks with 12 mL SD agar for seven days at 37 • C, spores were harvested through shaking with 5 mm glass beads after addition of 3 mL of saline. Spore suspensions containing approximately 10 6 spores/mL in sterile distilled water were used in the microprocessor controlled Autoplate Spiral Plating System AP5000 (Advanced Instruments, Norwood, MA, USA) according to the manufacturer's instructions as described previously [17]. Depending on the molecular weight of the compounds, an up to a 200-fold fungicide dilution gradient on SD agar was achieved using 15 cm plates. The concentration ranges (µg/mL) for the different fungicides were: boscalid (0.1-18.469), carbendazim (0.1-11.464), imazalil (0. 25-43.153), itraconazole (0.025-6.113 or 0.1-22.324), posaconazole (0.005-1.0) (pyraclostrobin (0.1-20.120), tebuconazole (0.1-17.349), terbinafine (0.01-1.7) and voriconazole (0.1-19.120). The different concentration ranges were chosen to distinguish sensitive wild-type (wt) isolates without known resistance mechanisms from those of insensitive isolates harboring resistance mechanism(s) in a single assay. Isolates were streaked on the spiral SD agar plates (eight per plate) from the outside to the center using cotton swaps and incubated at 37 • C in the dark. After 24 h incubation, the fungal growth of each isolate on the spiral plate was visually assessed and MIC values determined using the Spiral Gradient Endpoint (SGE) software.

DNA Extractions
After harvesting spores from one-week cultures in tissue culture flasks, 1.5 mL of spore suspensions was transferred into a 2 mL tube and centrifuged for 2 min at 13,200 rpm. After removing the supernatant, DNA was extracted according to the MasterPure Yeast DNA Purification kit (Lucigen Corporation, Madison, WI, USA) with the inclusion of an extra bead-beating step, which involved the addition of glass beads (0.425-0.600 mm) and the use of the Genie 2 Vortex (Scientific Industries, Bohemia, NY, USA) at full power for 2 min. This step was carried out after the lysis step just before adding the MPC Protein Precipitation Reagent.

PCR Amplification and Sequencing of Fungicide Resistant Alleles
Whole or partial fungicide target genes (β-tubulin, cytochrome b and succinate hydrogenase subunit B), covering all regions where mutations affecting inhibitor binding have been reported for fungi, were amplified with PCR, sequenced, and analyzed using Geneious software version 10.0 (Biomatters, Auckland, New Zealand) according to the previously established protocols [17].

Cell Surface Protein Typing
The cell surface protein (CSP) (XM_749624.1) encoding gene of Af was also partially sequenced from selected strains [29]. CSP typing of strains, including CSP t02* or t02B [17], was carried out according to the established nomenclature [26], which is based on the tandem repeat region, in which up to ten different 12-bp repeat sequences have been found in different copy numbers, and the flanking regions. After manual alignment of sequences, a phylogenetic tree for the different CSP sequences encountered in this study was constructed using the Geneious Tree Builder software (Biomatters, Auckland, New Zealand) based on the Tamura-Nei distance model and the Neighbor-Joining method.

Microsatellite Typing Based on Short Tandem Repeats in A. fumigatus (STRAf)
In total, 128 Af isolates with different levels of azole sensitivity from this study and our previous studies on environmental and reference isolates were further characterized using microsatellite genotyping based on a panel of nine short tandem repeat markers (STRAf 2A, 2B, 2C, 3A, 3B, 3C, 4A, 4B and 4C) according to the method previously described and validated [27,30]. GenAlEx v.6.503 software was used to identify distinct multi-locus genotypes (MLGs), where a single difference in an allele size was considered enough to discriminate a unique MLG [31]. To visualize the phylogenetic relationships among MLGs, a minimum spanning network (MSN) was generated using the R package poppr v.2.8.3 [32], which is based on Bruvo's distance [33].

Isolation and Fungicide Sensitivity Testing of A. fumigatus Isolates from Tulip Field Soils
In total, 180 isolates were isolated from soils sampled from six fields at five different locations (Table 1). Figures 1 and 2 show the sensitivity profiles for the overall population to eight different fungicides.
Only one azole sensitive strain, STNL1-A8 with wild-type CYP51A, was resistant to carbendazim. Several strains showing low levels of insensitivity to tebuconazole, in particular, carried F46Y/M172V/E427K. Amongst these strains, STNL2-C9 showed a raised MIC value of 6.017 µg/mL to pyraclostrobin. As expected, only small differences in sensitivity to terbinafine, a squalene epoxidase inhibitor, were measured for all isolates and this was not linked to a CYP51A variant. Of the medical azoles tested, posaconazole was most effective with the majority of isolates displaying MICs between 0.007 and 0.046 µg/mL, followed by voriconazole (most strains in MIC range 0.100 to 0.657 µg/mL), imazalil (most strains in MIC range 0.250 to 1.643 µg/mL), tebuconazole (most strains in MIC range 0.110 to 1.710 µg/mL) and itraconazole (most strains in MIC range 0.403 to 1.774 µg/mL), respectively ( Figure 1). A wide range of sensitivities were measured for itraconazole with a large fraction of isolates (n = 33), the majority carrying wild-type CYP51A alleles, having MIC values exceeding 10.000 µg/mL. MIC range 0.110 to 1.710 µg/mL) and itraconazole (most strains in MIC range 0.403 to 1.774 µg/mL), respectively ( Figure 1). A wide range of sensitivities were measured for itraconazole with a large fraction of isolates (n = 33), the majority carrying wild-type CYP51A alleles, having MIC values exceeding 10.000 µg/mL.  In total, 200 strains were isolated from tulip bulbs (n = 128), daffodil bulbs (n = 30) and tulip peel waste (n = 42) on SD agar without addition of tebuconazole and 19 strains

Azole Resistance Phenotype-to-Genotype Relationship, Cell Surface Protein and Mating Typing of A. fumigatus Isolates from the Environment with a Focus on Tulip Cultivation
A selection of 30 environmental Af isolates with different levels of insensitivity to azoles, carbendazim and/or pyraclostrobin were further tested for sensitivity to boscalid and were further characterized by determination of CSP, mating type and fungicide resistant alleles ( Table 2).

Isolation and Fungicide Sensitivity Testing of A. fumigatus Isolates from Flower Bulbs, Tulip Peel Waste Heaps and Compost
In total, 200 strains were isolated from tulip bulbs (n = 128), daffodil bulbs (n = 30) and tulip peel waste (n = 42) on SD agar without addition of tebuconazole and 19 strains from compost (n = 19) on tebuconazole-amended SD agar (Table 1). Eight additional strains were also isolated from three different tulip peel waste samples using tebuconazole-amended SD agar. The results of the fungicide sensitivity tests for the four largest populations together with the tulip soil population are displayed in Figure 2, while test results for a selection of individual strains are shown in Table 2. As expected, 19 compost isolates (Table 1), all isolated from tebuconazole-amended SD agar, showed high levels of insensitivity to all azoles tested, with 18, 17 and 14 of these also having MIC values exceeding 17.349, 19.120 and 43.153 µg/mL, for tebuconazole, voriconazole and imazalil, respectively ( Figure 2). All 19 isolates were also highly insensitive to both carbendazim and pyraclostrobin having MIC values greater than 11.464 and 20.120 µg/mL, respectively.
Tulip peel waste heaps also contained high frequencies of multi-fungicide insensitive isolates, with 23 out of 42 strains tested insensitive to multiple azoles and pyraclostrobin, respectively ( Figure 2). Of these, 22 were also highly insensitive to carbendazim. Six out of 23 were highly insensitive to voriconazole, tebuconazole and imazalil showing MICs greater than 10 ppm (including TP UT4A-1 and TP UT4B-2), while the remaining isolates (including TP UT1B-1, TP UT1A-2 and TP UT5C-5), showed raised levels of insensitivity to one or more azoles. All 19 isolates that were sensitive to pyraclostrobin were also sensitive to all azoles tested with MICs lower than 1.0, 2.0 and 4.0 µg/mL for voriconazole, imazalil and tebuconazole, respectively. All eight strains isolated from tebuconazole-amended SD agar were insensitive to both carbendazim and pyraclostrobin and showed high levels of insensitivity to multiple azoles (including TP TEB5C-5).

Azole
Resistance Phenotype-to-Genotype Relationship, Cell Surface Protein and Mating Typing of A. fumigatus Isolates from the Environment with a Focus on Tulip Cultivation A selection of 30 environmental Af isolates with different levels of insensitivity to azoles, carbendazim and/or pyraclostrobin were further tested for sensitivity to boscalid and were further characterized by determination of CSP, mating type and fungicide resistant alleles ( Table 2).
β-tubulin sequence analysis revealed that insensitivity to carbendazim was conferred in all environmental isolates by the amino acid substitution F200Y (codon change TTC to TAC). Isolate STNL5-C5, carrying TR 34 /L98H/S297T/F495I, showed a double nucleotide change resulting in F200Y (TTC to TAT). High levels of pyraclostrobin insensitivity (MIC values > 20.120 µg/mL) were associated with the cytochrome b alteration G143A (GGT to GCT), whereas the lower level of insensitivity to pyraclostrobin (MIC value of 6.017 µg/mL) in isolate STNL2-C9 was linked to cytochrome b F129L (TTC to TTA). Insensitivity to boscalid was conferred by SdhB alteration H270Y (CAC to TAC) in STNL6-B1 and TC TEB6B-1.

Azole Resistance Phenotype-to-Genotype Relationship, Cell Surface Protein and Mating Typing of A. fumigatus Isolates from the Clinical Setting
Results for the panel of 20 clinical Af isolates, including AF65 and AF293 as reference isolates, are presented in Table 3. The order of the voriconazole sensitivity mirrored the results for the environmental isoles with the reference isolates AF65 (wild-type CYP51A) and AF293 (F46Y/M172V/N284T/D255E/E427K) most sensitive, followed by isolates carrying TR 34 /L98H/S297T/F495I, TR 34 /L98H and TR 46 /Y121F/T289A, respectively. With regard to imazalil, four isolates with TR 34 /L98H/S297T/F495I showed a higher level of insensitivity, approximately three-fold, than most TR 34 /L98H isolates. In comparison with the MIC value of the wild-type isolate AF65, the average MIC value of the TR 34 /L98H/S297T/F495I isolates was approximately 15-fold higher (Resistance Factor (RF) of 15). TR 46 /Y121F/T289A were most insensitive, with RF > 25. TR 34 /L98H and TR 34 /L98H/S297T/F495I isolates showed different levels of insensitivity to tebuconazole (MIC values ranging from 3.341 to >17.349 µg/mL) with several strains also highly insensitive to carbendazim (MIC values > 11.464 µg/mL). Two TR 34 /L98H isolates, ARAF017 and CYP_15_46, were insensitive to fungicides belonging to all four different modes of action (azoles, MBC and QoI and SDHI fungicides). All six TR 46 /Y121F/T289A isolates showed high levels of insensitivity to tebuconazole and were also higly insensitive to both carbendazim (MIC values > 11.464 µg/mL) and pyraclostrobin (MIC values > 20.120 µg/mL). One isolate, CYP_15_46, was also insensitive to boscalid (MIC value > 18.469 µg/mL). β-tubulin sequence analysis revealed that insensitivity to carbendazim was conferred in all but one isolate by the amino acid substitution F200Y (codon change TTC to TAC). Carbendazim insensitivity in isolate CYP_15_46 was conferred by the beta-tubulin alteration E198A (GAG to GCG). Like the enviromental strains, mutations resulting in cytochrome b G143A (GGT to GCT) and SdhB H270Y (CAC to TAC) were found in all isolates insensitive to carbendazim and pyraclostrobin, respectively.

Genetic Diversity of Azole-Resistant Aspergillus fumigatus Isolated from the Clinical Setting and the Wider Environment with a Focus on Tulip Cultivation
In total, 128 Af isolates were analysed using STRAf typing. These originated from air (n = 5) and arable soils (n = 43) as part of our previous studies, and from tulip field soils (n = 13), tulip bulbs (n = 29), tulip peel waste (n = 13), tulip compost (n = 1) and patients (n = 18) as described here. Six reference strains that were used in both studies were also included. Most isolates were also further characterised with CSP and mating gene typing as well as CYP51A sequence analysis (see Table S1 for details). The STRAf profile for tulip compost isolate TC TEB6B-1 was inconclusive, showing two different product sizes for six out of the nine loci, indicative of a mixed culture, and was excluded for further analysis. Among 127 isolates, 99 distinct unique multilocus genotypes (MLGs) were identified, of which 88 were represented by single isolates (Figure 3).  Nodes with multiple isolates (clonal lineages), clusters, are arranged alphabetically. Nodes that are more closely related have darker and thicker edges, whereas nodes that are more distantly related have lighter and thinner edges. Names of a selection of isolates, including all medical strains, are also displayed and all details on strains can be found in Table S1.
Eleven MLGs were detected in multiple strains (clusters A to K), which can be regarded as clonal lineages. Cluster I was the largest cluster with ten isolates, all carrying TR46/Y121F/T289A, CSP t01 and MAT1-1, and insensitive to both carbendazim and pyraclostrobin ( Table 2). This cluster included seven strains isolated from tulip bulbs originating from the Netherlands and the UK (T3-5, T4-9, T4-10, T5-1, T5-2, T5-5 and T7-9), two strains from tulip waste heaps (including TP UT4A-1) and reference strain, TR46-NL, isolated from the environment in the Netherlands and tested previously [17]. Cluster C consisted of nine isolates carrying TR34/L98H, CSP t11 and MAT1-1, with no insensitivity observed for carbendazim and pyraclostrobin. Eight isolates originated from The relatedness between MLGs is based on Bruvo's genetic distances, which accounts for the stepwise mutation of microsatellite loci. Each node represents an MLG with one or more individuals. Nodes with multiple isolates (clonal lineages), clusters, are arranged alphabetically. Nodes that are more closely related have darker and thicker edges, whereas nodes that are more distantly related have lighter and thinner edges. Names of a selection of isolates, including all medical strains, are also displayed and all details on strains can be found in Table S1.

Discussion
Tulip bulbs have previously been identified as a vehicle for the international spread of azole resistant Af isolates [23,24], and decaying flower bulb waste from farms has already been identified as a hotspot in a previous study [25]. However, little is known about the emergence and spatiotemporal spread of different fungicide resistant alleles belonging to different modes of action, such as MBC, QoI and SDHI fungicides [17], in this environment. In this study, we sampled Af populations from tulip field soils, tulip bulbs, tulip peel waste heaps and tulip grower compost and compared their fungicide sensitivity status to both azole and non-azole fungicides. A selection of fungicide insensitive isolates associated with tulip cultivation and a panel of medical isolates were further characterized for presence of fungicide resistant alleles and genotyped using CSP and STRAf typing [26,27]. Several clones were identified and the level of clonal expansion of pan-azole and non-azole resistant strains in both the clinical setting and the wider environment is also further discussed hereinafter.
The 5.6% frequency (ten out of 180 isolates) of pan-azole insensitive Af strains isolated from Dutch tulip field soils in 2016 (Table 1) was higher than the reported 0.5% (two out 428 isolates) measured for arable soils sampled at nine different locations in France, Germany, and the UK in 2015 using the same sampling and test procedures [17]. Three out of the ten pan-azole insensitive isolates, all sampled from the same tulip field carried TR 46 /Y121F/T289A, while six TR 34 /L98H and one TR 34 /L98H/S297T/F495I strain were identified as well. One TR 34 /L98H, STNL6-B1, and all three TR 46 /Y121F/T289A isolates, STNL5-B7, STNL5-C1 and STNL5-C7, were also insensitive to both the MBC fungicide carbendazim and the QoI fungicide pyraclostrobin ( Table 2) which was attributed to alterations in β-tubulin (F200Y) and cytochrome b (G143A), respectively, as previously reported [17]. One of the TR 46 /Y121F/T289A strains, STNL5-C8, shared a MLG (26-21-9-33-11-22-8-14-10) with clinical isolates from the Netherlands and environmental isolates from India and Tanzania [35,36]. Interestingly, STNL6-B1 was also insensitive to the SDHI fungicide boscalid and carried the SdhB alteration H270Y, a change associated with SDHI resistance [17]. The TR 34 /L98H/S297T/F495I isolate, STNL5-C5, was also insensitive to carbendazim and showed a double mutation resulting in the F200Y β-tubulin alteration. One F46Y/M172V/E427K isolate, STNL2-C9, showing low levels of insensitivity to azoles, was also moderately insensitive to pyraclostrobin (MIC value of 6.017 µg/mL) and carried cytochrome b F129L, an alteration well known to confer low to moderate levels of QoI resistance in plant pathogenic fungi [37], and recently also reported for Af [19]. Only one azole sensitive strain, STNL1-A8, carrying wild-type CYP51A appeared to have evolved carbendazim insensitivity by acquiring the beta-tubulin F200Y allele, which has been reported once before [19]. Azole resistance seems to have spread quickly after the emergence of TR 34 -and TR 46 -based CYP51A variants, with subsequent selection within these populations for resistance to fungicides belonging to other modes of actions in environments where multiple classes of fungicides are present. The presence of an azole sensitive isolate with carbendazim insensitivity can be explained by the rare emergence and survival of this strain in an environment with and without MBC fungicides, in the absence of azole fungicides or before azoles were introduced. Alternatively, this isolate could also be the offspring from a cross between a multi-fungicide resistant azole insensitive isolate and a wild-type strain.
High numbers of Af colonies (up to 100 per bulb) were isolated from the tulip bulb samples when sampling the outer peels, which is in accordance with other studies investigating tulip and flower bulb samples [23,24]. The frequency of pan-azole resistant isolates was higher for the population isolated from tulip bulbs with 25 out of 108 isolates (23.1%) detected in 2015. The frequency was higher for bulbs purchased in the Netherlands (36.7%, 22 out of 60) than for bulbs purchased in the UK (6.3%, three out of 48), but the frequency for the UK-produced bulb population could be lower as it is likely that some purchased bulbs were directly imported from the Netherlands and not cultivated in the UK. No pan-azole resistant isolates were detected in populations sampled from Dutch tulip bulbs or UK daffodil bulbs in 2017, although the population size was smaller with 20 and 30 isolates, respectively. Seven out of the 25 pan-azole resistant isolates from the tulip bulb population carried TR 46 /Y121F/T289A and were also highly insensitive to both carbendazim and pyraclostrobin. These seven isolates originated from different tulip bulb samples purchased in the Netherlands and the UK in 2015 and formed together with two Dutch tulip peel waste heap strains, isolated in 2018, and a Dutch environmental reference strain (TR46-NL), isolated before 2014, a cluster (cluster I in Figure 3) with an identical CSP type (t01), mating type (MAT1-1) and MLG (26-21-12-26-9-20-13.3-9-9), indicative for a clonal lineage [28]. The same MLG clone has also been widely reported for both clinical and environmental strains from the Netherlands in the past [35,36], showing its spread and longevity through asexual reproduction. The successful expansion of this clone can be explained by a greater fitness in comparison with other genotypes. The presence of both β-tubulin F200Y and cytochrome b G143A alleles, conferring resistance to MBC and QoI fungicides, respectively, will increase the chance of survival in environments where multiple modes of fungicides are used (bulb dipping, seed treatments and foliar sprays) or can be found (plant waste and compost heaps). One pan-azole resistant tulip bulb isolate carrying TR 34 /L98H (T11-8) was also insensitive to carbendazim and pyraclostrobin and can also be considered as a clonal lineage as strains with an identical MLG (14-20-9-31-9-10-8-10-28) have been reported in clinical and environmental strains from India and in environmental strains from Tanzania [36,38]. The successful rapid expansion of this clone in India can be explained by its multi-fungicide resistance status, the ability to outcompete other strains in the presence of MBC and/or QoI fungicides upon its arrival in India. However, the emergence of this strain in India through sexual recombination and subsequent spread to other continents cannot be ruled out [38]. In this context it would be interesting to see if the clonal expansion of a TR 34 /L98H strain (MLG 22-10-9-9-9-23-8-10-8) reported in compost in 13 different cities in Iran can also be explained by the presence of additional fungicide resistant alleles in this clone [39]. The remaining 17 pan-azole resistant isolates were all sensitive to both carbendazim and pyraclostrobin. At least 11 of these carried TR 34 /L98H, from which at least eight were grouped together with a 2016 sugar beet field soil isolate from Belgium in cluster C, having an identical MLG (20-21-16-77-12-11-16.3-11-21), CSP t11 and MAT1-1 in common. As far as we know, this clone has not been reported before.
As expected, the highest frequency of pan-azole resistant isolates, 52.4%, was detected in samples collected from tulip peel waste heaps with 22 out of 42 isolates detected. The frequency of pan-azole resistant strains in compost samples could not be determined as Af isolates were mainly recovered from tebuconazole-amended agar due to excessive growth of other fungi on agar without addition of tebuconazole. High frequencies of up to 21.5% for recovery of pan-azole resistant strains from flower bulb waste heaps has also been reported in earlier studies [25]. Almost all pan-azole resistant strains (21 out of 22) were also highly insensitive to both carbendazim and pyraclostrobin, non-azole fungicides with different modes of action that are commonly found in tulip peel or flower waste heaps because of their application in bulb dipping and/or as part of foliar sprays to control diseases like Fusarium and Botrytis. Further evaluation of tulip peel waste heap and compost isolates showed that most multi-fungicide resistant isolates with high levels of voriconazole insensitivity (MIC value > 19.120 µg/mL) carried TR 46 /Y121F/T289A, but a more complex CYP51A variant, TR 34 /L98H/T289A/I364V/G448S, was also detected in two isolates that were isolated from colonies growing on tebuconazole-amended agar ( Table 2). This variant has been found in clinical strains from South Korea [40] and the UK since 2016 [41], but has also been reported in strains isolated from imported Dutch tulip bulbs in Japan in 2018 [24], and from compost heap samples in the UK in 2019 that were sampled as part of a citizen-science project [42]. The emergence of more complex variants over time shows similarities with the evolution of azole resistance in the plant pathogen Zymoseptoria tritici, where a stepwise accumulation of CYP51 mutations has resulted in adaptation to higher azole dose rates and/or new azoles with different binding properties entering the market [43].
The collection of clinical isolates used in this study showed a wide variety of fungicide sensitivity pheno-and genotypes that were also found in the environmental isolates. CYP51A variants TR 34 /L98H, TR 34 /L98H/S297T/F495I and TR 46 /Y121F/T289A, were detected in eight, four, eight and six isolates, respectively. The highest levels of insensitivity to voriconazole, imazalil and tebuconazole were generally measured for isolates carrying TR 46 /Y121F/T289A, while TR 34 /L98H/S297T/F495I isolates were generally less sensitive to imazalil than TR 34 /L98H isolates which can be explained by the presence of F495I, which particularly affects the binding of imidazoles such as prochloraz and imazalil [12]. Three out of the four TR 34 /L98H/S297T/F495I isolates carried beta-tubulin F200Y and were highly insensitive to the MBC fungicide carbendazim. All six TR 46 /Y121F/T289A strains carried both beta-tubulin F200Y and cytochrome b G143A alleles, the latter conferring high levels of insensitivity to the QoI fungicide pyraclostrobin. One TR 46 /Y121F/T289A strain, CYP_15_38 with a rare CSP type t09, was also insensitive to boscalid, a SDHI fungicide, and carried SdhB H270Y. Two out of eight TR 34 /L98H isolates, ARAF017 and CYP_51_46, were also insensitive to all three non-azole fungicides carbendazim, pyraclostrobin and boscalid. One new fungicide resistant allele was detected, beta-tubulin E198A in isolate CYP_51_46, conferring resistance to carbendazim. This allele, like beta-tubulin F200Y, commonly found in plant pathogens that developed resistance to MBC fungicides [44], has recently been reported in Af [17].
Finally, tulip peel waste heap isolate TP UT5C-5 and clinical isolate OKH50, isolated in Japan in 2016, shared the same MLG (14-21-8-31-9-6-8-10-20) (cluster F in Figure 3) with an Irish clinical strain D5 isolated in 2015 [23], all carrying TR 34 /L98H. Traits like CSP t02 and MAT1-2 were shared between TP UT5C-5 and OKH50, but only TP UT5C-5 was highly insensitive to pyraclostrobin. This shows that pan-azole resistant clones of Af can spread and develop de novo resistant alleles under exposure to different fungicides in hotspots. Using MLGs only based on STRAf typing might underestimate the spread of clones as low levels of instability have recently been reported for markers 3A and 3C. The recently reported MLG (14-21-8-32-9-6-8-10-20) for Chinese farm soil isolates carrying TR 34 /L98H could also represent descendants of the same clone as OKH50 [45]. Whole genome sequencing and/or further typing using additional markers targeting hypervariable TRs within exons of surface protein coding genes will be needed to confirm this [41,46].

Conclusions
The understanding of the emergence and epidemics of pan-azole resistant Af over time has been improved by this and other recently published research linking strains found in patients with genotypes distributed in the wider environment [41,47]. The acquisition of azole-resistant alleles in clinical and/or environmental settings, as well as non-azole fungicide resistant alleles in the environment, have contributed to the rapid expansion of clonal lineages in hotspots under selection by fungicides, where sexual reproduction can also generate new genotypes with combinations of fungicide resistant alleles. Continuous real-time surveillance of aerosol samples at different locations for the presence of Af combined with identification of CYP51A variants and other fungicide resistant alleles using a NGS genomics approach can provide information on the emergence and spatiotemporal dynamics of newly evolved strains. The detection of newly evolved azole resistant strains and clones can also be used to identify hotspots, where measures can be devised to reduce reproduction and/or fungicide selection pressure, aiming to minimize the risk of human exposure to airborne pan-azole resistant Af conidia through inhalation. Strains carrying novel CYP51A variants, based on new mutations and or combinations of (new) mutations, as well as strains harboring different azole resistance mechanisms [48], should be tested against all available azole antifungals, as changes in their azole susceptibility status can guide decisions for an optimal diagnosis and management of aspergillosis in patients [49].