The Molecular Identification and Antifungal Susceptibility of Clinical Isolates of Aspergillus Section Flavi from Three French Hospitals

(1) Background: Aspergillus flavus is a cosmopolitan mold with medical, veterinary, and agronomic concerns. Its morphological similarity to other cryptic species of the Flavi section requires molecular identification techniques that are not routinely performed. For clinical isolates of Aspergillus section Flavi, we present the molecular identification, susceptibility to six antifungal agents, and clinical context of source patients. (2) Methods: One hundred forty fungal clinical isolates were included in the study. These isolates, recovered over a 15-year period (2001–2015), were identified based on their morphological characteristics as belonging to section Flavi. After the subculture, sequencing of a part of the β-tubulin and calmodulin genes was performed, and resistance to azole antifungals was screened on agar plates containing itraconazole and voriconazole. Minimum inhibitory concentrations were determined for 120 isolates by the European Committee on Antimicrobial Susceptibility Testing (EUCAST) broth microdilution method. (3) Results: Partial β-tubulin and calmodulin sequences analysis showed that 138/140 isolates were A. flavus sensu stricto, 1 isolate was A. parasiticus/sojae, and 1 was A. nomiae. Many of the isolates came from samples collected in the context of respiratory tract colonization. Among probable or proven aspergillosis, respiratory infections were the most frequent, followed by ENT infections. Antifungal susceptibility testing was available for isolates (n = 120, all A. flavus ss) from one hospital. The MIC range (geometric mean MIC) in mg/L was 0.5–8 (0.77), 0.5–8 (1.03), 0.125–2 (0.25), 0.03–2 (0.22), 0.25–8 (1.91), and 0.03–0.125 (0.061) for voriconazole, isavuconazole, itraconazole, posaconazole, amphotericin B, and caspofungin, respectively. Two (1.67%) isolates showed resistance to isavuconazole according to current EUCAST breakpoints with MICs at 8 mg/L for isavuconazole and voriconazole. One of these two isolates was also resistant to itraconazole with MIC at 2 mg/L. (4) Conclusions: The present characterization of a large collection of Aspergillus belonging to the Flavi section confirmed that A. flavus ss is the predominant species. It is mainly implicated in respiratory and ENT infections. The emergence of resistance highlights the need to perform susceptibility tests on section Flavi isolates.

A. flavus is also of great importance due to its ability to cause disease in animals and crops and to produce carcinogenic mycotoxins [36].Another epidemiological characteristic of A. flavus is its higher prevalence in tropical countries [37][38][39].
Nevertheless, the prevalence of the different cryptic species in clinical samples is largely unknown, as is their pathogenicity.None of the cryptic species in section Flavi are known to carry natural resistance to antifungal drugs [46].However, A. flavus shows high amphotericin B MICs compared to A. fumigatus [47].
In this study, we molecularly analyze 140 isolates belonging to the Flavi section and determine the susceptibility profile for a subset (n = 120) of the isolates.

Study Design, Patients and Isolates
A total of 140 clinical isolates of Aspergillus section Flavi recovered from 107 patients were analyzed.Isolates were mainly cultured from respiratory samples (n = 114) and ENT sphere (n = 19).Seven isolates were cultured from other sites: two artery biopsies, dialysis fluid, nail, arm biopsy, hallux biopsy, and endocardial biopsy (Table S1).All isolates were collected between 2001 and 2015 through routine clinical work from patients of three hospitals: Hôpital Henri Mondor (HMN), Hôpital Universitaire de Lile (LIL), and Hôpital Européen Georges Pompidou (HEGP).The number of isolates collected each year was homogeneous over time, with 72 isolates between 2001 and 2008 and 68 isolates between 2009 and 2015.These isolates were stored frozen.Patients' identifiable information had already been anonymized.Since the study was conducted on isolates collected through routine clinical work and patients' identifiable information had already been anonymized, no written or verbal informed consent was necessary for patients to participate in this study.We testify that we followed the ethical standards of the Helsinki Declaration of 1975, as revised in 2008.Patients include lung transplant patients with cystic fibrosis and other pathologies, patients with hematological malignancies, and patients with ENT diseases and other pathologies.Clinical data showed that most of the isolates (96 isolates) were considered colonizers; 41 isolates were considered implicated in infections, including invasive pulmonary aspergillosis (12 isolates), otitis (12 isolates), bronchopulmonary aspergillosis (6 isolates), sinusitis (1 isolate), aspergilloma (1 isolate), endovascular infection (2 isolates), endocarditis (1 isolate), nasal aspergillosis (1 isolate), hallux infection (1 isolate), arm infection (1 isolate), and one disseminated invasive aspergillosis (1 isolate).For five isolates, clinical data were not available.

Molecular Identification
DNA extraction: Molds were subcultured in Sabouraud medium for 7 days at 35 • C. Conidia were subjected to a mechanical shock with MagNA Lyser Green Beads (Roche Diagnostics, Meylan, France) in MagNA Lyser Instrument (Roche) and thermal shock in the ice to disrupt the wall.Then, after 4 h incubation at 56 • C with proteinase K (Qiagen Sciences Inc., Courtaboeuf, France), DNA was extracted using the QIAamp DNA Blood Mini Kit (Qiagen Sciences Inc.) following the manufacturer's instructions.
Amplification was performed on a Gene Amp ® PCR System 9700 (Applied Biosystems; Thermo Fisher, Waltham, MA, USA).The PCR conditions were initial denaturation at 95 • C for 10 min, followed by 35 cycles of denaturation at 95 • C for 30 s, annealing at 55 • C for 30 s, extension at 72 • C for 1 min, and final extension at 72 • C for 10 min.
After purification through columns of the MinElute PCR Purification Kit (Qiagen Sciences Ing.), sequencing was performed by Sanger's method (IMRB Genomics Platform).Sequences were analyzed using Bioedit (Hall, T.A. Version 7.2.5).Each sequence from our isolates was compared to the sequence of the type strains of each species within the Flavi section.Sequences used for comparison are based on the most recent taxonomic revision of the Flavi section published by Houbraken et al. in 2020 [41].These sequences are presented in Table S2.Isolates were identified to a given species if the homologies with sequences of type strains were >98%.When there were mismatches between calmodulin and betatubulin, we used the most discriminating gene (between beta-tubulin and calmodulin) as defined by Houbraken et al. [41].Phylogenetic trees of β-tubulin and calmodulin genes were built with maximum parsimony likelihood method with MEGA (Version 11.0.13)software.β-tubulin and calmodulin genes partial sequences were concatenated, and phylogenetic trees were also generated with these concatenated sequences.Using a heatmap, we labeled the tree obtained from the concatenation of calmodulin and β-tubulin with the MICs for each isolate.

Antifungal Susceptibility Testing
Isolates were revived from storage and subcultured on Sabouraud-agar slanted tubes.In vitro resistance to azoles was screened by subculturing each isolate on RPMI agar plates supplemented with itraconazole (Sigma-Aldrich, Saint-Quentin Fallavier, France) at 4 mg/L and voriconazole (Sigma-Aldrich) at 1 mg/L, as previously described [59].After that, MICs for azoles were determined by EUCAST method.
The antifungal agents used in this study were amphotericin B (Sigma-Aldrich, Saint Quentin-Fallavier, France), itraconazole (Sigma-Aldrich), voriconazole (Pfizer Inc., New-York, NY, USA), isavuconazole (Basilea, Bâle, Suisse), posaconazole (Merck & Co Inc., Kenilworth, NJ, USA), and caspofungin (Merck & Co Inc.).The final concentrations tested ranged from 0.015 to 8 mg/L for each drug.Plates were incubated at 35 • C for 48 h in a humidified atmosphere.After incubation, the microplates were read with a reading mirror (for all drugs but caspofungin) to visualize the fungal growth in each well and determine the MIC as the lowest concentration with no visible growth.For caspofungin, an inverted microscope was used to determine the antifungal concentration that produced a visible change in the morphology of the hyphae compared with the growth control well (Minimum Effective Concentration).The EUCAST has set breakpoints for the interpretation of antifungal susceptibility testing results for itraconazole (Resistant isolates: MIC > 1 mg/L) and isavuconazole (Resistant isolates: MIC > 2 mg/L) [60].For other antifungal drugs, breakpoints are not available, and we used ECOFFs for the interpretation: isolates were considered non-wild type when MIC was >4 mg/L for amphotericin B; >2 mg/L for voriconazole and >0.5 mg/L for posaconazole.Breakpoints and ECOFFs of echinocandins have not been set yet, and rates of resistance have not been calculated.A. flavus ATCC 204304, Candida krusei ATCC 6258, and Candida parapsilosis ATCC 22019 were used as quality control strains.

Molecular Identification and Cryptic Species
Sequence analysis showed that 138 isolates were A. flavus ss, and two isolates were different cryptic species, A. parasisticus/sojae and A. nomiae.A. flavus ss sequences have been deposited in GenBank with accession numbers from OR285555 to OR285692 for beta-tubulin and accession numbers from OR226374 to OR266455 and from OR285498 to OR2885554 for calmodulin.For β-tubulin sequences, the phylogenetic trees built with 458 bp alignment length are shown in Figure S2.For calmodulin sequences, the phylogenetic trees built with 512 bp alignment length are shown in Figure S3.Among A. flavus ss isolates, 5 and 10 different sequence patterns were identified for β-tubulin and calmodulin, respectively (Tables S3 and S4).
The β-tubulin sequence (GenBank accession number OR455454) of isolate HEGP 1350 showed 99.7% and 97% homology with the reference sequences of A. nomiae and A. pseudonomiae, respectively.The calmodulin sequence (GenBank accession number OR455456) of isolate HEGP 1350 showed 98.6% and 99.1% homology with A. nomiae and pseudonomiae reference sequences, respectively.We conclude that this isolate is A. nomiae.
The β-tubulin sequence (GenBank access number OR455455) of isolate HEGP 3223 showed 99.2% and 100% homology with the reference sequences of A. parasiticus and A. sojae, respectively.The calmodulin sequence (GenBank access number OR455457) of isolate HEGP 3223 showed 100% homology with the reference sequences of A. parasiticus and A. sojae, respectively.
The phylogenetic tree of concatenated calmodulin and β-tubulin sequences is presented in Figure 1.Concatenated calmodulin and β-tubulin sequences showed 15 patterns (Table S5).The evolutionary history was inferred using the Neighbor-joining method [61].The optimal tree is shown.The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches [62].The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.The evolutionary distances were computed using the Maximum Composite Likelihood method [63] and are in the units of the number of base substitutions per site.This analysis involved 175 nucleotide sequences.All ambiguous positions were removed for each sequence pair (pairwise deletion option).There was a total of 970 positions in the final dataset.Evolutionary analyses were conducted in MEGA11 [64].

Antifungal Susceptibility Testing and Azole MICs Distribution
For logistical reasons, we were only able to determine MICs on 120 isolates from the HEGP hospital.Resistance screening revealed two voriconazole-resistant isolates collected in December 2005 and September 2007 among the 120 A. flavus ss isolates.
Tables 1 and 2 show the distribution and values of MICs for the 120 A. flavus ss determined by the EUCAST method.For A. flavus ss isolates, the geometric mean MICs of voriconazole, isavuconazole, itraconazole, posaconazole, amphotericin B, and caspofungin were 0.78, 1.04, 0.25, 0.22, 1.90, and 0.06 mg/L, respectively.Of the 120 A. flavus ss isolates, 2 (1.67%) had isavuconazole MICs higher than clinical breakpoints.These two isolates were cross-resistant to voriconazole and isavuconazole with MICs of 8 mg/L for both drugs and were non-wild type for posaconazole with MICs of 1 mg/L.One of these two isolates had an itraconazole MIC of 2 mg/L, which can be interpreted as resistance to this drug.Of the remaining isolates, six others had posaconazole MICs of 1 mg/L, and one isolate had an amphotericin MIC of 8 mg/L.Table 3 shows MICs for A. nomiae and A. parasiticus/sojae.Geometric mean MICs were 0.50, 0.71, 0.35, 0.18, 2.83, and 0.04 mg/L for voriconazole, isavuconazole, itraconazole, posaconazole, amphotericin B, and caspofungin, respectively.The evolutionary history was inferred using the Neighbor-joining method [61].The optimal tree is shown.The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches [62].The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.The evolutionary distances were computed using the Maximum Composite Likelihood method [63] and are in the units of the number of base substitutions per site.This analysis involved 175 nucleotide sequences.All ambiguous positions were removed for each sequence pair (pairwise deletion option).There was a total of 970 positions in the final dataset.Evolutionary analyses were conducted in MEGA11 [64].

Antifungal Susceptibility Testing and Azole MICs Distribution
For logistical reasons, we were only able to determine MICs on 120 isolates from the HEGP hospital.Resistance screening revealed two voriconazole-resistant isolates collected in December 2005 and September 2007 among the 120 A. flavus ss isolates.
Tables 1 and 2 show the distribution and values of MICs for the 120 A. flavus ss determined by the EUCAST method.For A. flavus ss isolates, the geometric mean MICs of voriconazole, isavuconazole, itraconazole, posaconazole, amphotericin B, and caspofungin were 0.78, 1.04, 0.25, 0.22, 1.90, and 0.06 mg/L, respectively.Of the 120 A. flavus ss isolates, 2 (1.67%) had isavuconazole MICs higher than clinical breakpoints.These two isolates were cross-resistant to voriconazole and isavuconazole with MICs of 8 mg/L for both drugs and were non-wild type for posaconazole with MICs of 1 mg/L.One of these two isolates had an itraconazole MIC of 2 mg/L, which can be interpreted as resistance to this drug.Of the remaining isolates, six others had posaconazole MICs of 1 mg/L, and one isolate had an amphotericin MIC of 8 mg/L.Table 3 shows MICs for A. nomiae and A. parasiticus/sojae.Geometric mean MICs were 0.50, 0.71, 0.35, 0.18, 2.83, and 0.04 mg/L for voriconazole, isavuconazole, itraconazole, posaconazole, amphotericin B, and caspofungin, respectively.

Discussion
In the present study, we report on the precise identification of a large collection of Flavi section isolates and their antifungal susceptibility.Given their importance in human, animal, and plant pathologies and their use in food and pharmaceutical industries, a better knowledge of the species of section Flavi is of prime importance.According to the last taxonomical revision, the Flavi section comprises 35 species grouped in eight series [41].In the present study, most of the isolates were A. flavus ss (98.4%), showing that cryptic species seem to be rare in clinical samples.Within A. flavus ss isolates, calmodulin polymorphism was more marked than beta-tubulin polymorphism, as also observed by Frisvad [40].The former led to the definition of 10 clusters, while the latter revealed 5 clusters.Concatenation of the two sequences revealed 15 clusters.Overlaying these clusters with susceptibility data revealed no specific profile for any given cluster.Based on calmodulin and beta-tubulin partial sequences, only two cryptic species were identified: A. parasiticus/sojae and A. nomiae.As the epidemiology of Aspergillus may depend on climatic conditions and the type of patients hospitalized, our results do not necessarily reflect the overall epidemiology in hospitals in France.Despite the mismatch between calmodulin and beta-tubulin for isolate HEGP 1350, since beta-tubulin is discriminating between A. nomiae and A. pseudonomiae, whereas calmodulin is not [40], we concluded that our isolate is A. nomiae.Neither beta-tubulin nor calmodulin discriminate between A. parasiticus and A. sojae.A. sojae is a domesticated nontoxigenic species associated with A. parasiticus, as A. oryzae is associated with A. flavus [40,65].Several cases of infection due to cryptic species of section Flavi have been reported in the literature (Table 4).Some cases of A. flavus ss infection reported in France, with underlying disease, treatment, and outcome, are presented in Table S5. A. parasiticus and A. nomiae are two of the major aflatoxin-producing species.A. nomiae has been reported in cases of pulmonary disease [46,[66][67][68], onychomycosis [69], keratitis [6,70], and rhinofacial aspergillosis [71].In our study, both species were recovered from pulmonary specimens but were considered colonizers.Among the other cryptic species, A. pseudonomiae has been reported in a sinus tissue sample from a patient with fungal rhinosinusitis [72] and in keratitis [73]. A. alliaceus, A. caelatus, A. minisclerotigenes, A. novoparasiticus, A. tamarii, A. pseudotamarii, and A. oryzae var.effusus are the other cryptic species that have been isolated from clinical samples.A. tamarii has been reported as a cause of primary cutaneous aspergillosis [74,75], wound infection [72,76,77], onychomycosis [78], keratitis [6][7][8][9][10]79], and invasive nasosinusal aspergillosis [80,81].A. pseudotamarii could also cause keratitis [11].A. alliaceus is an etiological agent of invasive pulmonary aspergillosis [46,82,83], and this species produces ochratoxin [84]. A. oryzae var.effusus has been reported as a cause of keratitis [6] and has been isolated from the lower respiratory tract of hospitalized patients [85].In 2019, one case of A. caelatus airway colonization in a chronic obstructive pulmonary disease patient was reported in Colombia [86]. A. minisclerotigenes has been reported as a cause of sinus [45] infection and keratitis [87]. A. novoparasiticus has been isolated from the sputum of a patient in São Paulo, Brazil [88].The low number of clinical cases could be explained by an underestimation of cryptic species due to their difficulty in identification.Although molecular sequencing of β-tubulin/calmodulin remains the gold standard for a precise species identification, new technologies such as whole-genome sequencing are now used for taxonomic purposes [89].Other molecular techniques are also available for assessing the genetic diversity among A. flavus isolates [90].Mass spectrometry may be interesting in routine clinical microbiology laboratories [91].Environment plays an important role in the epidemiology of A. flavus human infection.Infections due to this species are more prevalent in dry and hot climatic regions, and some clinical cases reported in France are imported infections contracted in Africa [24,25,32].Infections are mostly acquired by exposure to A. flavus spores in the environment, and nosocomial infections in hospitals have also been reported from surgery unit air and surface [30,92].Among our isolates, two resistant isolates (1.67%) with voriconazole and isavuconazole MICs of 8 mg/L recovered from patients with long-term treatment with voriconazole were detected.This observation is in favor of the selection for cross-resistance to voriconazole and isavuconazole by antifungal pressure during the patient's treatment with voriconazole.The first clinical isolate of voriconazole-resistant A. flavus has been recovered from the lung biopsy of a Chinese patient who had also previously been treated with voriconazole [53].Voriconazole-resistant isolates have been reported from China [53], India [54,55,93], South Korea [56], and Spain [94,95], with a prevalence between 2.5 and 14% among clinical isolates.One of our isolates showed amphotericin MIC at 8 mg/L.Cases of resistance to amphotericin B have also been reported, and the MICs of amphotericin B for A. flavus are generally higher than for A. fumigatus [96].In the present study, none of the azoleresistant isolates and isolates with MICs of 8 mg/L for amphotericin B were involved in invasive aspergillosis.
The emergence of resistant isolates highlights the importance of antifungal susceptibility testing, particularly for patients undergoing long-term treatment.The best techniques are the microdilution broth reference techniques (EUCAST or CLSI), but other techniques, such as the concentration gradient strip (CGS) method, can also be used.Indeed, it has been shown that the results obtained with the CGS method correlated well with those obtained with the reference techniques for A. flavus [97].

Conclusions
This study demonstrated that cryptic species within the Flavi section and azoleresistant A. flavus ss can be found among clinical isolates.Therefore, precise identification to the species level and routine antifungal susceptibility testing are needed.

Figure 1 .
Figure 1.Phylogenetic tree of A. flavus of concatenated partial β-tubulin and calmodulin sequences.(A) All isolates and reference sequences.(B) Only A. flavus ss with A. aflatoxiformans as outgroup.The evolutionary history was inferred using the Neighbor-joining method[61].The optimal tree is shown.The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches[62].The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.The evolutionary distances were computed using the Maximum Composite Likelihood method[63] and are in the units of the number of base substitutions per site.This analysis involved 175 nucleotide sequences.All ambiguous positions were removed for each sequence pair (pairwise deletion option).There was a total of 970 positions in the final dataset.Evolutionary analyses were conducted in MEGA11[64].

Figure 1 .
Figure 1.Phylogenetic tree of A. flavus of concatenated partial β-tubulin and calmodulin sequences.(A) All isolates and reference sequences.(B) Only A. flavus ss with A. aflatoxiformans as outgroup.The evolutionary history was inferred using the Neighbor-joining method[61].The optimal tree is shown.The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches[62].The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.The evolutionary distances were computed using the Maximum Composite Likelihood method[63] and are in the units of the number of base substitutions per site.This analysis involved 175 nucleotide sequences.All ambiguous positions were removed for each sequence pair (pairwise deletion option).There was a total of 970 positions in the final dataset.Evolutionary analyses were conducted in MEGA11[64].

Table 1 .
MIC distribution of 6 antifungal drugs against 120 A. flavus ss isolates.

Table 2 .
Statistical indices of the distribution of MICs for 6 antifungal drugs against 120 A. flavus ss isolates.

Table 3 .
MICs for the two cryptic species.

Table 4 .
Flavi section cryptic species in the literature.