Species Distribution and In Vitro Azole Susceptibility of Aspergillus Section Nigri Isolates from Clinical and Environmental Settings

Abstract

Aspergillus section Nigri includes species of interest for animal and human health, although studies on species distribution are limited to human cases. Data on the antifungal susceptibilities and the molecular mechanism of triazole resistance in strains belonging to this section are scant. Forty-two black Aspergillus strains from human patients (16 isolates), animals (14 isolates), and the environment (12 isolates) were molecularly characterized and their in vitro triazole susceptibilities investigated. Aspergillus tubingensis was isolated from humans, animals, and environmental settings, whereas Aspergillus awamori and Aspergillus niger were isolated exclusively from humans. Phylogenetic analyses of β-tubulin and calmodulin gene sequences were concordant in differentiating A. tubingensis from A. awamori and A. niger. Voriconazole and posaconazole (PSZ) were the most active triazoles. One A. tubingensis strain was resistant to itraconazole and PSZ and one A. niger strain to PSZ. Sequence analysis of the cyp51A gene revealed different sequence types within a species, and A. tubingensis strains were also phylogenetically distinct from A. awamori/A. niger strains according to the strain origin and susceptibility profile. Genetic analysis of the cyp51A sequences suggests that two nonsynonymous mutations resulting in amino acid substitutions in the CYP51A protein (changes of L to R at position 21 [L21R] and of Q to R at position 228 [Q228R]) might be involved in azole resistance. Though azole resistance in black Aspergillus isolates from animals and rural environments does not represent a threat to public health in Southern Italy, the use of triazoles in the clinical setting needs to better monitored. The cyp51A sequence is useful for the molecular identification of black Aspergillus, and point mutations in protein sequences could be responsible for azole resistance phenomena.

INTRODUCTION

Aspergillus section Nigri includes 19 species of black Aspergillus, which are widely known for being able to cause food spoilage and to produce enzymes and organic acids used in the fermentation industry, as well as mycotoxins (1). Nevertheless, Aspergillus section Nigri includes species causing pulmonary aspergillosis and otomycosis in humans, as well as localized and disseminated disease in domestic and wild animals (2, 3). Species belonging to Aspergillus section Nigri have been difficult to classify due to their phenotypic similarities (4), whereas molecular tools such as the sequencing of calmodulin and β-tubulin genes have been successfully employed for species identification and discrimination within section Nigri. Molecular data have shown that Aspergillus tubingensis is the species most frequently distributed in various environments (5–7). The management and prophylaxis of aspergillosis is mainly performed using triazole drugs, although long-term therapy and the indiscriminate use of azoles in agriculture have raised concerns about resistance to these compounds (8–10). In spite of the large number of reports of triazole resistance phenomena worldwide, mainly in Aspergillus fumigatus isolates from environmental and clinical settings (9, 10), data on the in vitro antifungal susceptibilities of members of Aspergillus section Nigri are scant and those available show that clinical isolates from different geographical regions exhibited remarkably different azole susceptibilities (5, 7, 11, 12). Azole resistance mechanisms have been described mainly in A. tubingensis (5), the most common being attributed to point mutations in the cyp51A gene encoding the azole target protein (14α-sterol demethylase [CYP51A]) (9, 10). However, the molecular mechanisms of azole resistance in species of Aspergillus section Nigri are controversial and not well defined (5).

In light of this situation, accurate species identification of black Aspergillus isolates and their susceptibility testing are instrumental to assess the occurrence and molecular mechanisms of azole resistance phenomena and fungal species distribution among different clinical settings and environments. Thus, the aims of this study were to (i) identify the species of black Aspergillus isolates from different clinical and environmental sources based on sequence analysis of the calmodulin and β-tubulin genes, (ii) evaluate their in vitro susceptibilities to triazoles, and (iii) investigate the potential mechanisms of azole resistance.

MATERIALS AND METHODS

Isolates.

A total of 42 black Aspergillus isolates were tested in the study. The isolates came from 567 human patients, 125 squirrels, and 57 air samples from 19 sheds from laying hen farms, as follows. Sixteen clinical isolates were from human patients with hematological disorders (n = 5), cystic fibrosis (n = 6), or lung disease (n = 5) admitted to different wards of Azienda Ospedaliera-Universitaria, Ospedale Policlinico Consorziale di Bari, between September 2015 and March 2016. Fourteen isolates were from the respiratory tract of squirrels (Callosciurus finlaysonii; Basilicata, Southern Italy) trapped during a campaign to control allochthonous wildlife populations (13). Twelve isolates were from the environment of laying hen farms (14).

All animal and human cases of probable/possible invasive aspergillosis (IA) were defined according to the criteria of the European Organization for Research and Treatment of Cancer/Mycoses Study Group (EORTC/MSG) (15). All isolates were morphologically identified as belonging to Aspergillus section Nigri using macro- and microscopic features (16). The strains were stored at −80°C in the fungal collection of the Department of Veterinary Medicine at the University of Bari, Italy. Prior to testing, each isolate was subcultured at least twice onto Sabouraud dextrose agar (SDA) plates to ensure purity and viability.

Molecular identification and phylogenetic analysis.

DNA was extracted from isolates grown on SDA for 5 days at 35°C by using the ArchivePure DNA yeast kit (5-Prime, Inc., USA) according to the manufacturer's instructions.

The isolates were molecularly identified by PCR and sequencing of partial calmodulin and β-tubulin genes using the primer pair Cmd5 (5′-CCGAGTACAAGGAGGCCTTC-3′) and Cmd6 (5′-CCGATAGAGGTCATAACGTGG-3′) and the primer pair BT2a (5′-GGTAACCAAATCGGTGCTGCTTTC-3′) and BT2b (5′-ACCCTCAGTGTAGTGACCCTTGGC-3′), respectively (17, 18). The thermocycling conditions were an initial denaturation at 95°C for 7 min, followed by 35 cycles of 95°C for 30 s, 61°C for 30 s (for the β-tubulin gene) or 57°C for 30 s (for the calmodulin gene), and 72°C for 1 min, followed by a final extension step of 72°C for 10 min.

The resultant PCR amplicons were purified using the ExoI/FAST AP enzyme systems (Thermo Scientific). Purified PCR products were sequenced using the TaqDyeDoxy terminator cycle sequencing kit (version 2; Applied Biosystems) in an automated sequencer (ABI Prism 377). Sequences were aligned using the ClustalX program and compared with those available in the GenBank database by using the Basic Local Alignment Search Tool (BLAST; http://blast.ncbi.nlm.nih.gov/Blast.cgi).

Phylogenetic analysis was performed using the calmodulin and β-tubulin gene sequences of all black Aspergillus isolates. Phylogenetic trees were produced from ClustalX-aligned sequences by the maximum-likelihood (ML) method (Kimura 2-parameter model) (19) and neighbor-joining (NJ) (20), based on the p-distance of each gene (calmodulin and β-tubulin). Evolutionary distances were computed using the software package MEGA 6.0 (Center for Evolutionary Medicine and Informatics, Tempe, AZ). The reliability of internal branches was assessed using the bootstrap method with 1,000 replicates, and bootstrap values of >50 were considered significant in this analysis. Aspergillus carbonarius was chosen as the outgroup in both sequence analyses.

In vitro antifungal susceptibility.

In vitro susceptibility testing of black Aspergillus isolates to determine their MICs of itraconazole (ITZ), voriconazole (VRZ), and posaconazole (PSZ) was performed by the broth microdilution (BMD) method as described in CLSI document M38-A2 (21). The triazole MICs were determined as described by the CLSI reference method (21). In vitro susceptibility testing was performed in triplicate for each isolate. The proposed epidemiological cutoff values (ECVs) for ITZ (2 mg/liter), VRZ (2 mg/liter), and PSZ (0.5 mg/liter) were used for the interpretation of results (22). Quality control was performed as recommended in the M38-A2 CLSI document using Candida krusei (strain ATCC 6258) and Candida parapsilosis (strain ATCC 22019) (American Type Culture Collection, Manassas, VA, USA). The MIC data obtained were reported as the ranges, mean values, and MIC₅₀ and MIC₉₀.

Sequencing of cyp51A.

The cyp51A sequences were obtained for 24 isolates (16 clinical human isolates, 4 from squirrels, and 4 from laying hen farms).

The partial cyp51A gene was amplified by PCR using the primers Ancyp51A-F (5′-TKTYCCTGCCTACRGTCGCTT-3′) and Ancyp51A-R (5′-CCGTAGTCCACCATCTCTCC-3′) (5). The thermal cycling profile for amplification was as follows: 94°C for 5 min, followed by 45 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 2 min, with a final step of 72°C for 10 min. Sequencing and phylogenetic analysis were conducted as described above.

Phylogenetic analysis was performed using the cyp51A sequences of 24 black Aspergillus isolates from this study and those available in the GenBank database.

Nucleotide sequence accession numbers.

All representative sequence types of A. tubingensis, Aspergillus niger, and Aspergillus awamori have been deposited in the GenBank database (for β-tubulin, accession numbers KX231820 to KX231822, and for calmodulin, accession numbers KX231823 to KX231827). The GenBank accession numbers for representative cyp51A sequences are KX245375 to KX245382.

RESULTS

Isolates.

The clinical data, diagnoses, antifungal prophylaxis or therapy, and clinical outcomes of the 16 patients that scored positive for Aspergillus spp. are summarized in Table 1. In particular, Aspergillus species were isolated from patients who suffered from hematological disorders (n = 5) and cystic fibrosis (n = 6) with high and low respiratory tract fungal colonization, respectively, and from patients with probable IA (n = 5) (Table 1). Aspergillus spp. were isolated from squirrels with possible IA (n = 2) or colonization of the nasal cavity (n = 12).

TABLE 1.

Clinical data, diagnosis, antifungal prophylaxis or therapy, and clinical outcome of the 16 patients that scored positive for Aspergillus spp.

Patient code	Gender	Age (yr)	Underlying diseasea	Diagnosisb	Antifungal prophylaxis/therapyc	Clinical outcome	Strain code	Aspergillus sp.	Sample typed	Azole susceptibilitye			Nonsynonymous mutation in cyp51A
										ITZ	PSZ	VRZ
ET16	F	51	AML	RTC	None	Alive	CD1450	A. tubingensis	NS	S	S	S	None
ET21	F	37	ALL	RTC	None	Alive	CD1445	A. tubingensis	NS	S	S	S	None
ET19	F	66	MM	RTC	P: FLZ 400 mg/day for 10 days	Alive	CD1443	A. tubingensis	PS	S	S	S	None
ET22	M	42	AML	RTC	P: PSZ 600 mg/day for 15 days	Alive	CD1199	A. tubingensis	NS	R	R	S	L21F
ET23	F	50	ALL	RTC	P: PSZ 600 mg/day for 18 days	Alive	CD1200	A. niger	NS	S	R	S	R228Q
FC5	F	20	CF	RTC	None	Alive	CD1461	A. awamori	Sp	S	S	S	None
FC13	F	21	CF	RTC	None	Alive	CD1452	A. awamori	Sp	S	S	S	None
FC15	F	38	CF	RTC	None	Alive	CD1455	A. niger	Sp	S	S	S	None
FC16	M	11	CF	RTC	None	Alive	CD1456	A. tubingensis	Sp	S	S	S	None
FC21	M	5	CF	RTC	None	Alive	CD1498	A. awamori	Sp	S	S	S	None
FC	F	33	CF	RTC	None	Alive	CD1201	A. tubingensis	Sp	S	S	S	None
R8	F	40	Lung disease	PIA	T: VRZ 200 mg/day	Alive	CD1202	A. awamori	Sp	S	S	S	None
R9	M	61	Lung disease	PIA	T: VRZ 200 mg/day	Alive	CD1497	A. awamori	Sp	S	S	S	None
R10	M	56	Lung disease	PIA	T: VRZ 200 mg/day	Alive	CD1463	A. awamori	Sp	S	S	S	None
A8	M	70	Lung disease	PIA	T: VRZ 200 mg/day	Alive	CD1449	A. niger	Sp	S	S	S	None
A9	M	59	Lung disease	PIA	T: PSZ 800 mg/day	Died	CD1462	A. awamori	Sp	S	S	S	None

^aAML, acute myeloid leukemia; ALL, acute lymphatic leukemia; MM, multiple myeloma; CF, cystic fibrosis.

^bRTC, respiratory tract colonization; PIA, probable invasive aspergillosis.

^cP, prophylaxis; T, therapy; FLZ, fluconazole; VRZ, voriconazole; PSZ, posaconazole.

^dNS, nasal swab; PS, pharyngeal swab; Sp, sputum.

^eITZ, itraconazole; S, susceptible; R, resistant.

Molecular identification and phylogenetic analysis.

The PCR amplification of individual DNA samples resulted in amplicons of the expected size (∼550 bp for the β-tubulin gene and ∼580 bp for the calmodulin gene). Three distinct species were identified among all 42 samples according to the β-tubulin and calmodulin gene sequence analysis. All β-tubulin and calmodulin gene sequences displayed a 99 to 100% nucleotide identity with sequences available in GenBank (Table 1). Interestingly, A. tubingensis, the most common Aspergillus species identified (76.2%) (Table 1), was the only species isolated from human patients, animals, and environmental settings, whereas A. awamori and A. niger were isolated exclusively from human patients.

Pairwise comparisons revealed a single sequence type (ST) for β-tubulin genes within each species, herein designated At_βT for A. tubingensis, An_βT for A. niger, and Aa_βT for A. awamori. Three calmodulin STs were detected among A. tubingensis strains (herein designated At_C1, At_C2, and At_C3), with nucleotide variation ranging from 0.2 to 0.3%. A single ST was identified among A. niger and A. awamori isolates (herein designated An_C1 and Aa_C1, respectively). There was consistency in the topology of the trees inferred by the ML and NJ methods (for each locus; data not shown), and the phylogenetic analyses of β-tubulin and calmodulin sequences were concordant in grouping the 42 isolates into two different clades, dividing A. tubingensis isolates from A. niger/A. awamori (Fig. 1 and 2). In the calmodulin tree, three sister subclades grouped A. tubingensis STs (At_C1, At_C2, and At_C3), and within the clade there is support (bootstrap value of >50) for clustering of isolates according to their origin (squirrels, humans, and laying hen farms) (Fig. 2).

FIG 1.

Maximum-parsimony tree based on partial β-tubulin sequences. The laboratory code is reported for each strain, and the sequence type is indicated in parentheses.

FIG 2.

Maximum-parsimony tree based on partial calmodulin sequences. The laboratory code is reported for each strain, and the strain origin and sequence type, in parentheses, are indicated.

In vitro antifungal susceptibility.

The antifungal susceptibilities of the 42 isolates in this study are shown in Table 2. VRZ and PSZ were the triazoles most active against the three Aspergillus species tested. Higher MICs of ITZ were recorded in A. tubingensis than in the other Aspergillus species. Only two Aspergillus strains revealed azole resistance phenomena, and the corresponding data on the isolates and patients are summarized in Table 1. In particular, one strain (strain CD1199) of A. tubingensis was resistant to ITZ (MIC of 64 μg/ml) and PSZ (MIC of 1 μg/ml) and one strain of A. niger (strain CD1200) to PSZ (MIC of 1 μg/ml). These two strains were isolated from two patients admitted to the hematology unit with nasal colonization.

TABLE 2.

Numbers and percentages of Aspergillus sp. isolates obtained from human patients, squirrels, and laying hen farms

Aspergillus sp.	No. (%) of isolates from:
	Human patients	Squirrels	Laying hen farms	Total
A. tubingensis	6 (14.3)	14 (33.3)	12 (28.6)	32 (76.2)
A. awamori	7 (16.7)			7 (16.7)
A. niger	3 (7.1)			3 (7.1)
Total	16 (38.1)	14 (33.3)	12 (28.6)	42 (100)

cyp51A sequencing and phylogenetic analysis.

The sequence analysis of the partial cyp51A gene (∼1,230 bp) revealed 3 distinct species with 3 or 2 STs within a single species (Fig. 3). In particular, three cyp51A STs were detected among A. tubingensis (At_cy1, At_cy2, and At_cy3) and A. awamori (Aa_cy1, Aa_cy2, and Aa_cy3), respectively, and two STs were detected for A. niger (An_cy1 and An_cy2). Pairwise comparison among the different STs revealed nucleotide variation ranging from 1.3 to 2.8% in A. tubingensis isolates and from 0.3 to 0.9% in A. awamori isolates and nucleotide variation of 0.3% in A. niger isolates (Table 3).

FIG 3.

Maximum-parsimony tree based on partial cyp51A sequences. The laboratory code is reported for each strain, and sequence types are shown.

TABLE 3.

Itraconazole, voriconazole, and posaconazole MIC data for 42 Aspergillus species isolates

Aspergillus sp. (no. of isolates)	MIC (μg/ml) value(s) fora:
	ITZ				VRZ				PSZ
	Range	Mean (SD)	MIC50	MIC90	Range	Mean (SD)	MIC50	MIC90	Range	Mean (SD)	MIC50	MIC90
A. tubingensis (32)	0.25–64	3 (11.14)	1	2	0.5–2	0.69 (0.38)	0.5	1	0.25–1	0.49 (0.12)	0.5	0.5
A. awamori (7)	0.5–2	1 (0.5)	1	1	0.25–1	0.46 (0.27)	0.5	0.5	0.25–0.5	0.36 (0.27)	0.5	0.5
A. niger (3)	0.25–2	1.08 (0.88)			0.25–2	0.91 (0.94)			0.25–1	0.58 (0.38)

^aITZ, itraconazole; VRZ, voriconazole; PSZ, posaconazole.

In A. tubingensis, three nonsynonymous amino acid substitutions (changes of A to V at position 9 [A9V], T to A at position 321 [T321A], and L to F at position 21 [L21F]) were detected in the At_cy1 ST and were characteristic of the ITZ- and PSZ-resistant strain (strain CD1199). In particular, the A9V and T321A mutations were detected in strain CD1199 and in the reference resistant strain (strain NRRL4700), as well as in A. niger/A. awamori susceptible strains (strains CD1449, CD1461, and CD1498), whereas the L21F mutation was present only in strain CD1199. The At_cy2 ST, having only the T321A mutation, characterized the susceptible strains collected from laying hen farms. The At_cy3 ST was associated with susceptible strains originating from human patients and animals. In A. niger, the An_cy1 ST, characterizing the PSZ-resistant strain (strain CD1200), differed from the ST in the susceptible strains (ST An_cy2) by a nonsynonymous mutation (R228Q). The R228Q mutation has also been detected in reference resistant strains (strains NRRL341 and AN186). In A. awamori, all three STs (Aa_cy1, Aa_cy2, and Aa_cy3) characterized the susceptible strains and only the Aa_cy1 ST showed a nonsynonymous variation (F343L). The F343L mutation has not been detected in the reference resistant strain (strain F7577), which was characterized by the nonsynonymous mutation K97T.

Phylogenetic analysis of the partial cyp51A gene sequences revealed concordance with the β-tubulin and calmodulin sequences in grouping Aspergillus spp. into two clades, differentiating A. tubingensis from A. awamori/A. niger. The cyp51A data separated the strains according to their azole susceptibility profiles within the species. In particular, A. tubingensis strain CD1199 (At_cy1) clustered separately from the susceptible strains (Fig. 3) and from the resistant strain NRRL4700. Similarly, in the A. niger subclade, the azole-resistant (CD1200, NRRL341, and AN186) and azole-susceptible (ATCC 1015, CD1455, and CD1449) strains clustered separately (Fig. 3), and the A. awamori susceptible strains clustered separately from the corresponding resistant strain (strain F7577).

DISCUSSION

This study presents, for the first time, an overview of the occurrence of black Aspergillus from clinical and environmental sources in Southern Italy and a molecular characterization at the species level by sequencing the β-tubulin and calmodulin genes. In addition, to our knowledge, this is the first study that has characterized black Aspergillus from animals (squirrels) and rural environments (laying hen farms). In particular, A. tubingensis is one of the most prevalent species isolated from human patients (5, 11, 23, 24), but it was never isolated previously from animals and the environment. Albeit the sampling of squirrels and rural environments performed in this study was limited, the presence of A. awamori and A. niger was recorded only in human patients (5 with probable IA, 4 who suffered from cystic fibrosis, and 1 with a hematological disorder), thus indicating that these species might circulate in nosocomial environments and be encountered mainly in cystic fibrosis patients, as previously reported (5, 23, 24). Phylogenetic analyses based on β-tubulin and calmodulin gene sequences grouped the Aspergillus species consistently, indicating that these genes are good targets for molecular characterization of black Aspergillus at the species level (5). However, the calmodulin sequences were more suitable for molecular epidemiologic studies, since they allowed segregation of A. tubingensis strains according to their origin.

High VRZ and PSZ activities against Aspergillus spp. were confirmed (25–28) and are reported here for the first time in strains from animals and rural environments. A. tubingensis revealed lower susceptibility to ITZ than to VRZ and PSZ, as previously suggested (2, 29), and azole-resistant and cross-resistant isolates were detected only in strains from human patients (5). Indeed, in this study, A. tubingensis and A. niger isolates with low susceptibilities to ITZ and/or PSZ were isolated from hematological patients who received PSZ as prophylaxis in preventing IA. Therefore, the use of PSZ in hematopoietic stem cell transplant recipients might have induced or selected resistant strains in these patients (2, 29). The low prevalence of triazole resistance in black Aspergillus isolates (2.4%) detected in this study might be due to the low number of Aspergillus strains tested.

The finding of three nonsynonymous mutations (A9V, T321A, and L21F) in the CYP51A sequence of the A. tubingensis-resistant strain (CD1199) suggests a role in azole resistance or cross-resistance phenomena. Nonetheless, since the A9V and T321A mutations were previously detected both in the A. tubingensis resistant strain (NRRL4700) and in A. niger/A. awamori susceptible strains (5), further studies are necessary to elucidate their role in azole resistance mechanisms. Contrarily, the L21F mutation, detected only in the ITZ- and PSZ-resistant A. tubingensis strain CD1199, was reported here for the first time, suggesting its role in azole resistance or cross-resistance phenomena in this species. Likewise, PSZ resistance could also be caused by the nonsynonymous mutation Q228R that was detected only in the resistant A. niger strain CD1200 (see also strains NRRL341 and AN186 in Howard et al. 5). Since in A. awamori, the nonsynonymous mutation F343L was associated with susceptible strains, the mutation K97T, previously recorded in the reference resistant strain F7577 and lacking in the susceptible strains tested here, might play a role in azole resistance, as suggested by Howard and colleagues (5).

The phylogeny obtained with cyp51A largely reflected those obtained with the calmodulin and β-tubulin sequences, indicating that the gene may be useful as a taxonomic target to differentiate the Nigri clades, as previously reported (5). In that phylogeny, the two azole-resistant A. tubingensis and A. niger strains (CD1199 and CD1200) were clearly separated from the susceptible ones, thus indicating the relationship between their phylogeny and azole susceptibility profiles.

This study provides new data on the prevalence of black Aspergillus collected from different sources in Southern Italy and shows that the β-tubulin, calmodulin, and cyp51A genes were all useful for the molecular identification of this group of microorganisms, although the latter gene was more informative on strain origin and antifungal susceptibility.

The lack of recorded azole resistance in black Aspergillus isolates from animals and rural environments indicates that these sources do not currently represent a threat to public health in Southern Italy, whereas the use of triazoles in clinical settings might have induced or selected resistant strains; thus, in order to minimize the spread of these strains in nosocomial environments, as well as to ensure good management of Aspergillus species infections, the antifungal susceptibility profiles of isolates should be accurately monitored. Finally, not all of the amino acid substitutions identified in CYP51A protein sequences are responsible for azole resistance phenomena in black Aspergillus. However, the F21L and Q228R amino acid substitutions might have a role in triazole resistance of black Aspergillus and, since the number of isolates tested in this study is limited, confirmatory studies should be performed in order to support our data. Other molecular mechanisms of antifungal resistance, such as the overexpression of the cyp51A gene and of the multidrug efflux transporter genes, need to be investigated to clarify the molecular mechanisms involved in azole resistance of this group of microorganisms.

Funding Statement

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.