Comparison of Two Typing Methods for Characterization of Azole Resistance in Aspergillus fumigatus from Potting Soil Samples in a Chinese Hospital

To understand the characterizations of azole resistance in Aspergillus fumigatus from potting soil samples in the hospital, a total of 58 samples were collected. Among 106 A. fumigatus samples obtained, 5 isolates from 4 soil samples located in the gerontology department were identified as azole-resistant A. fumigatus (ARAF). Four ARAF isolates harbored the TR₃₄/L98H allele, and the other one had no mutation in the cyp51A gene.

ABSTRACT

To understand the characterizations of azole resistance in Aspergillus fumigatus from potting soil samples in the hospital, a total of 58 samples were collected. Among 106 A. fumigatus samples obtained, 5 isolates from 4 soil samples located in the gerontology department were identified as azole-resistant A. fumigatus (ARAF). Four ARAF isolates harbored the TR₃₄/L98H allele, and the other one had no mutation in the cyp51A gene. Among 174 A. fumigatus samples selected for genotyping, TRESPERG typing obtained a close discriminatory power (Simpson’s index of diversity [DI], 0.9941; 95% confidence interval [CI], 0.9913 to 0.9968) compared with the short tandem repeat (STR) typing (DI, 0.9997; 95% CI, 0.9976 to 1.0000). Genotyping showed that the TR₃₄/L98H isolates in the hospital had a close genetic relationship with ARAF isolates from China and many other countries. In conclusion, this study indicated the presence of ARAF in potting soil samples from the hospital, which might pose a risk of causing ARAF infection in patients.

INTRODUCTION

Aspergillus fumigatus is an opportunistic pathogen that is abundant in soil and decaying organic matter. Invasive fungal infections caused by Aspergillus represent a major threat for immunocompromised patients, such as those with solid organ transplantation, allogeneic hematopoietic stem cell transplantation, or congenital or acquired immune deficiency (1). Triazole antifungal agents are the main class of drugs used for treating aspergillus diseases. Since the first report of azole-resistant A. fumigatus (ARAF) in 1997 (2), there have been continuous reports of ARAF worldwide. A high rate of resistance has been reported in European countries such as the Netherlands (3), posing a great challenge for clinical treatment. Exposure to A. fumigatus, particularly ARAF, increases the risk of infection among hospitalized patients (4).

Flower bulb waste, green waste material, wood chippings, and compost (5, 6) were supposed to be the hotspots for breeding ARAF. A high prevalence of ARAF has also been reported in soil samples from urban city centers (7). However, the present studies were limited to flower beds nearby the hospital and in urban flower fields and public gardens (8, 9); no study has focused on the characterizations of azole resistance in A. fumigatus from the soil samples of flowerpots in hospitals. Although there was no evidence indicating that plants in hospitals pose an increased risk of infection, the Centers for Disease Control and Prevention (CDC) advises against growing flowers and plants in a protective environment.

In this study, we investigated the prevalence and characterization of ARAF from potting soil samples in different departments of the hospital and explored the genetic relationships of ARAF from different sources through two different typing methods.

RESULTS

Strain identification, antifungal susceptibility testing, and cyp51A mutation.

A total of 106 isolates from 58 flowerpots were identified as A. fumigatus sensu stricto through phenotypic and genotypic identification. After screening of azole resistance, eight environmental strains and one clinical strain showed growth on wells containing itraconazole, voriconazole, or posaconazole. Five environmental isolates from 4 soil samples were identified as ARAF according to the in vitro susceptibility test. All the four soil samples were collected from flowerpots of the gerontology department in the hospital. The prevalence of ARAF in the potting soil samples from the hospital was 6.9% (4 out of 58) (Table 1). Among the 5 ARAF isolates, 4 isolates harbored TR₃₄/L98H, while one isolate had no mutation in the cyp51A gene. All of the four A. fumigatus isolates with the TR₃₄/L98H allele exhibited a high level of MIC values to epoxiconazole, bromuconazole, tebuconazole, difenoconazole, and propiconazole. One clinical A. fumigatus isolate(C485) with the TR₃₄/L98H/S297T/F495I allele showed a high level of MIC values for all the seven azole fungicides (Table 2).

TABLE 1.

The prevalence of azole-resistant Aspergillus fumigatus (ARAF) isolates among soil samples from different departments

Department	Soil samples		Aspergillus fumigatus samples
	Total no.	No. of ARAF-positive samples (%)	Total no.	No. of ARAF isolates (%)
Interventional oncology	2	0 (0.0)	3	0 (0.0)
Hematology	3	0 (0.0)	5	0 (0.0)
Blood purification	4	0 (0.0)	9	0 (0.0)
Urology	1	0 (0.0)	0	(0.0)
Traditional Chinese medicine	2	0 (0.0)	8	0 (0.0)
Burns	2	0 (0.0)	12	0 (0.0)
Gerontology	10	4 (40.0)	15	5 (33.33)
Neurosurgery	2	0 (0.0)	2	0 (0.0)
Ophthalmology and otorhinolaryngology	1	0 (0.0)	0	0 (0.0)
Kidney internal medicine	4	0 (0.0)	11	0 (0.0)
Cardiology	5	0 (0.0)	9	0 (0.0)
Digestive	4	0 (0.0)	7	0 (0.0)
General medicine	1	0 (0.0)	0	0 (0.0)
Rehabilitation	4	0 (0.0)	9	0 (0.0)
Dermatology	6	0 (0.0)	4	0 (0.0)
Infectious diseases	2	0 (0.0)	3	0 (0.0)
Transfusion center	2	0 (0.0)	6	0 (0.0)
Flower beds inside the hospital	3	0 (0.0)	3	0 (0.0)
Total	58	4 (6.9)	106	5 (4.7)

TABLE 2.

Characteristics of 26 Aspergillus fumigatus isolates from the hospital

Strain	Source	Mating type	Azole resistance	Cyp51A mutation(s)	MIC (mg/liter)
					ITCa	VOR	POS	BRO	EPO	TEB	PRC	DIF	PRO	IMA
C475	Clinical	MAT1-2	S	None	0.5	1	0.25	4	8	4	1	2	8	0.5
C477	Clinical	MAT1-1	S	None	0.5	1	0.25	4	4	4	0.5	2	8	0.5
C478	Clinical	MAT1-1	S	None	1	0.5	0.13	2	4	4	1	4	8	0.5
C479	Clinical	MAT1-1	S	None	0.5	1	0.13	2	4	2	0.5	2	4	0.25
C481	Clinical	MAT1-1	S	None	0.5	0.5	0.06	1	2	1	0.25	1	2	0.13
C482	Clinical	MAT1-1	S	None	0.5	0.5	0.13	2	2	2	0.25	1	4	0.5
C483	Clinical	MAT1-1	S	None	0.5	0.5	0.13	1	2	2	0.25	2	4	0.25
C485	Clinical	MAT1-2	R	TR34/L98H/S297T/F495I	>16	2	1	>32	>32	32	>32	>32	>32	16
C487	Clinical	MAT1-2	S	None	0.5	1	0.13	2	2	2	0.5	2	4	0.5
C2089	Clinical	MAT1-1	S	None	1	0.5	0.25	2	4	4	1	4	4	0.25
C2090	Clinical	MAT1-2	S	None	0.5	0.25	0.5	1	4	2	0.5	2	2	0.25
C2091	Clinical	MAT1-1	S	None	0.5	0.5	0.25	2	4	4	1	4	8	0.5
C2092	Clinical	MAT1-1	S	None	0.5	0.25	0.25	2	4	4	0.5	2	4	0.25
C2093	Clinical	MAT1-2	S	None	1	0.5	0.25	4	8	4	0.5	4	8	0.5
C2094	Clinical	MAT1-2	S	None	0.5	0.5	0.25	2	2	2	0.25	1	2	0.13
C2095	Clinical	MAT1-1	S	None	1	0.5	0.25	2	4	2	0.25	2	4	0.25
E2041-2	Environmental	MAT1-2	R	None	>16	2	0.5	16	>32	>32	1	32	>32	2
E2042-1	Environmental	MAT1-2	R	TR34/L98H	>16	4	1	8	>32	16	0.5	8	32	1
E2045-1	Environmental	MAT1-1	R	TR34/L98H	>16	4	0.5	16	>32	>32	2	16	>32	4
E2045-2	Environmental	MAT1-1	R	TR34/L98H	>16	2	0.5	32	>32	>32	2	16	>32	2
E2048-2	Environmental	MAT1-1	R	TR34/L98H	>16	4	1	16	>32	32	1	16	>32	2
E2034-2	Environmental	MAT1-2	S	None	0.5	1	0.13	2	4	2	0.5	2	8	0.25
E2038-2	Environmental	MAT1-1	S	None	1	0.5	0.25	2	8	4	1	4	8	0.5
E2041-3	Environmental	MAT1-2	S	None	1	0.5	0.25	1	4	2	0.25	2	4	0.25
E2052-3	Environmental	MAT1-1	S	None	1	1	0.25	4	8	4	1	4	8	0.25
E2083-3	Environmental	MAT1-1	S	None	0.5	0.5	0.25	1	2	2	0.5	2	4	0.25

^aITC, itraconazole; VOR, voriconazole; POS, posaconazole; BRO, bromuconazole; EPO, epoxiconazole; TEB, tebuconazole; DIF, difenoconazole; PRO, propiconazole; IMA, imazalil; PRC, prochloraz; R, azole resistance; S, azole susceptibility.

Distribution of mating types.

The results of mating types showed that there were 3 MAT1-1 and 3 MAT1-2 azole-resistant A. fumigatus isolates from this hospital. There was no significant difference in the proportions of mating types between the clinical and environmental isolates (χ², 1.477; P, 0.224) and the ARAF and azole-susceptible A. fumigatus isolates (χ², 2.846; P, 0.092) among the 174 A. fumigatus isolates.

TRESPERG typing improvement.

After sequencing and analyzing of 4 alleles among the 174 A. fumigatus isolates by TRESPERG typing, we found that there was a mistake for the original alleles of CFEM, named c19, proposed by Garcia-Rubio et al. (10) after we analyzed the c19 sequence in GenBank. The correct order named c19 should be 06-01-01-01-01-02-01-03-03-04-03-05-04-03-03—04-03-03-03-06-01-03-09-10 (em dash indicates irregular sequence, which are not tandem repeat succession). Combined with previous studies, the nucleotide and amino acid sequence of each repeat and alleles for cell surface protein (CSP) typing are shown in Tables S1 and S2 in the supplemental material.

In total, the number of CSP, ERG4B, MP2, and CFEM genotypes were 17, 16, 20, and 23 in the 174 A. fumigatus isolates, respectively, of which 2 ERG4B, 6 MP2, and 6 CFEM genotypes have not been reported in previous studies. (Details of the variants and repeats are shown in Supplemental Tables S3 to S5.)

Discrimination of two different typing methods.

For the 174 A. fumigatus isolates analyzed, there were 125 different genotypes in TRESPERG typing and 162 different genotypes in short tandem repeat (STR) typing. The discriminatory power of TRESPERG typing (DI, 0.9941; 95% confidence interval [CI], 0.9913 to 0.9968) is close to that of STR typing (DI, 0.9997; 95% CI, 0.9976 to 1). The discriminatory power of CSP typing is relatively lower (DI, 0.8240; 95% CI, (0.7951 to 0.8529). The Simpson diversity index of the two typing methods and each locus are shown in Table 3.

TABLE 3.

Simpson’s index of diversity (DI) of TRESPERG and STR typing in 174 Aspergillus fumigatus isolates

Typing method or locus	No. of types	DI
TRESPERG typing	125	0.9941
CSP	17	0.8240
ERG4B	16	0.8338
MP2	20	0.7651
CFEM	23	0.8164
STR typing	162	0.9997
2A	16	0.9049
2B	23	0.9106
2C	20	0.9084
3A	46	0.9700
3B	24	0.8784
3C	30	0.9556
4A	20	0.8936
4B	16	0.8481
4C	17	0.8506

Genotype analysis and genetic relationships of A. fumigatus.

CSP typing showed that the three ARAF isolates with the TR₃₄/L98H allele correspond to type t02 and t11, the one ARAF isolate with no mutation in the cyp51A gene corresponds to type t06A, and the TR₃₄/L98H/S297T/F495I clinical isolate belongs to type t04A. The most frequent CSP types were t01 and t04A, which accounted for 29.89% (n = 52) and 20.11% (n = 35) of 174 A. fumigatus isolates. The typing results of the ARAF isolates in this hospital are shown in Table 4.

TABLE 4.

Genotyping results of azole-resistant Aspergillus fumigatus isolates in this hospital^a

Strain ID	TRESPERG typing				STR typing
	CSP	ERG4B	MP2	CFEM	2A	2B	2C	3A	3B	3C	4A	4B	4C
C485	t04A	e11	m3.4	c04	23	12	15	23	22	6	13	10	7
E2041-2	t06A	e05	m1.1	c09	14	10	12	25	9	9	12	10	17
E2042-1	t11	e05	m1.1	c09	14	21	8	30	9	7	8	10	10
E2045-1	t02	e15	m1.1	c08A	13	10	9	38	9	10	8	10	8
E2045-2	t02	e15	m1.1	c08A	13	10	9	38	9	10	8	10	8
E2048-2	t02	e15	m1.1	c08A	13	10	9	38	9	10	8	10	8

^aSTR 2A, 2B, 2C, 3A, 3B, 3C, 4A, 4B, and 4C are 9 microsatellite markers. CSP, ERG4B, MP2, and CFEM are 4 loci for TRESPERG typing.

The microscale thermophoresis (MST) analysis of TRESPERG showed that the four environmental ARAF isolates in this study were clustered in the lower clade, and the clinical ARAF C485 was in the upper clade. The MST analysis of STR typing showed that the four environmental ARAF isolates were located in the left upper clade of the MST. Both typing methods suggested the environmental ARAF isolates were genetically unrelated to the clinical ARAF. Three ARAF isolates (E2045-1, E2045-2, and E2048-2) shared the same genotype by both TRESPERG genotyping and STR genotyping. In TRESPERG genotyping, the three strains shared the same genotype with four strains, which were isolated from Shanghai (STJ0105) and Fuzhou (C135, C136) in China and from Zambia (E1934). The clinical ARAF (C485) shared the same genotype with an environmental isolate from Beijing (E739).

Results of repeat sampling.

A total of 78 A. fumigatus isolates were obtained from 18 samples, and 16 isolates were identified as ARAF (20.51%). Twelve ARAF isolates (75%) harbored the TR₃₄/L98H allele in the cyp51A gene. t11 was the most common CSP type in the 16 ARAF isolates (n = 12; 75%). Although CSP type t11 was commonly identified in European countries, it has not been reported in China (11). Meanwhile, a strain harboring the TR₃₄/L98H/S297T/F495I allele in the cyp51A gene was isolated. The CSP types for the other 4 ARAF isolates include t01, t06A, and t06B. These results suggest that there was diversity of ARAF in potting soil samples from the hospital.

DISCUSSION

Exposure to environmental ARAF increases the risk of infection for patients. Flower bulb waste and green waste material are considered as the hotspots for breeding ARAF (5). This study showed that A. fumigatus could be isolated from a majority of flowerpots. All of the five ARAF were isolated from the gerontology department, posing a threat to the patients admitted to the department.

Similarity in spatial molecule structure between agricultural demethylase inhibitors (DMIs) and medical triazoles results in cross-resistance of A. fumigatus to different azoles. Therefore, it is vital to understand the occurrence of azole resistance in different environments. The most common cyp51A mutation in this study was TR₃₄/L98H. A prospective multicenter international surveillance suggested that this allele originated from Italy and Austria (12). A. fumigatus carrying the TR₃₄/L98H/S297T/F495I allele usually causes resistance to itraconazole and leads to high MIC values to DMIs, for which the prevalence has been found to be high in China (13–17). In this study, we have obtained similar results; elevated MIC values to imazalil and prochloraz have been observed among the A. fumigatus isolates carrying the TR34/L98H/S297T/F495I allele (including strain E739 from our previous study [13]), but not those with the TR34/L98H allele.

We using ultrahigh-performance liquid chromatography coupled to tandem mass spectrometry (UHPLC-MS/MS) to detect the DMI residues (18). No residues of DMIs have been detected in the ARAF-positive samples in this study. It suggested that the ARAF in the flowerpots were derived from the hospital environment but not from the selection pressure of soil samples.

Many typing methods have been described for A. fumigatus, and each of them has its advantages and disadvantages. The methods based on the electrophoretic-specific bans were limited to poor repeatability. Other methods, such as multilocus sequence typing (MLST), have good repeatability but showed poor discriminatory power. CSP typing is widely used because of its simple operation and good stability, while STR typing is widely used because of its high discriminatory power. As a quick response method, CSP typing could be regarded as a good option to avoid additional genotyping in an urgent outbreak. In this case, TRESPERG was used, as it has a higher discriminatory power. The result of genotyping showed that three ARAF isolates shared the same genotype; all of them came from the wards and corridors of the gerontology department, suggesting that there was clonal spread of ARAF in this department.

In conclusion, our study provided evidence of ARAF in flowerpots in hospitals, which might pose a risk of causing ARAF infection in patients. We strongly recommend that plants shouldn’t be kept in hospitals, especially in wards with high-risk patients. Immunocompromised patients should be housed in a protective environment. (19, 20). The result of genotyping showed that TRESPERG typing is a cost-saving method with high discriminatory power.

MATERIALS AND METHODS

Collection of soil samples and isolation of A. fumigatus.

Fifty-eight soil samples were collected from different flowerpots in 18 departments of a tertiary care hospital in northeast China in October 2018. Two grams of each soil sample were dissolved in 8 ml saline containing 0.1% Tween 20; then they were fully dissolved by vortex shock (13). From this suspension, we placed 100 μl on the Sabouraud dextrose agar plates (SDA) supplemented with 50 mg/liter chloramphenicol and incubated at 42°C for 3 to 5 days. Three colonies, which were phenotypically suspected as A. fumigatus, were randomly selected from each plate for further identification; all of the colonies were selected if their numbers were less than three in the plate. All the suspected A. fumigatus isolates were finally identified by growth at 48°C and sequencing of the β-tubulin gene (21). In addition, we included 43 environmental strains, which were isolated from soil samples in China and Zambia, for comparison of genotyping.

Sources of clinical isolates.

A total of 7 clinical A. fumigatus samples isolated from the patients in 2018 were collected. Nine clinical strains from our previous study in this hospital (14) and 105 clinical strains, which originated in China, Belgium, and the Netherlands, were used for comparison of genotyping.

Screening of azole resistance in A. fumigatus.

All the A. fumigatus isolates were screened for azole resistance in this study using 4-well plates, which contained RPMI 1640 agars supplemented with itraconazole (4 mg/liter), voriconazole (2 mg/liter), posaconazole (0.5 mg/ml), and a drug-free growth control (22). The isolates that grew on the antifungal-free and on any of the drug-containing wells were considered as potential azole-nonsusceptible isolates.

Antifungal susceptibility testing and cyp51A gene sequencing.

Antifungal susceptibility was conducted for the 16 clinical isolates, all the suspected azole-nonsusceptible environmental isolates, and a subset of azole-susceptible environmental isolates according to the EUCAST broth microdilution method E.DEF 9.3 (13, 23). In this study, the drugs tested include 3 clinical azoles (itraconazole, voriconazole, posaconazole) and 7 sterol 14α-demethylase inhibitors (DMIs) (epoxiconazole, bromuconazole, tebuconazole, prochloraz, difenoconazole, propiconazole, and imazalil). All isolates identified as ARAF by the antifungal susceptibility test were subjected for sequencing of the cyp51A gene and its promoter region as previously described (24).

Genotyping of A. fumigatus.

Mating type (25), TRESPERG typing, and STR typing with nine microsatellite loci (STRAf 2A, 2B, 2C, 3A, 3B, 3C, 4A, 4B, and 4C) were determined for 174 A. fumigatus samples as previously described (10, 26, 27), including 26 isolates from this study (Fig. 1 ) and 148 randomly selected isolates collected in our laboratory (13, 14). The Simpson’s index of diversity (DI) was calculated for TRESPERG typing and STR typing to compare their discriminatory powers (28). The genetic relationship of clinical and environmental A. fumigatus from the hospital was analyzed by BioNumerics 7.5 through a categorical analysis of 9 microsatellite markers using the unweighted pair group method with arithmetic means (UPGMA) clustering. A minimum spanning tree (MST) was constructed based on TRESPERG typing of the 174 A. fumigatus isolates by BioNumerics 7.5 (Fig. 2).

FIG 1.

Genotypic relationships among 26 clinical and environmental Aspergillus fumigatus isolates in the hospital. The dendrogram is based on a categorical analysis of nine microsatellite markers in combination with UPGMA clustering. NA, not available.

FIG 2.

A minimum spanning tree (MST) showing the genotypic diversity of Aspergillus fumigatus isolates in the hospital. (A)STR MST, and (B) TRESPERG MST. Each circle shows a unique genotype, and the size indicates the number of this genotype. Connecting lines between circles show the similarities between genotypes. The thicker the line, the closer relationship they have. A thick solid line indicates difference in one marker, a thin solid line indicates differences in two or three markers, a dashed line indicates different in four markers, and the different markers above 4 are shown as a dotted line. The different colors of these circles indicate the different of cyp51A mutation types.

Resampling and isolation of A. fumigatus from the department where ARAF was isolated.

After the antifungal testing of A. fumigatus, a total of 18 potting soil samples were collected repeatedly from the departments where ARAF was isolated. A. fumigatus was isolated and analyzed using the same methods described above, except that the number of strains isolated from each sample increased from a maximum of three to a maximum of five.