Point Mutation or Overexpression of Aspergillus fumigatus cyp51B, Encoding Lanosterol 14α-Sterol Demethylase, Leads to Triazole Resistance

ABSTRACT

Aspergillus fumigatus is the most common cause of invasive fungal mold infections in immunocompromised individuals. Current antifungal treatment relies heavily on the triazole antifungals which inhibit fungal Erg11/Cyp51 activity and subsequent ergosterol biosynthesis. However, resistance, due primarily to cyp51 mutation, is rapidly increasing. A. fumigatus contains two Cyp51 isoenzymes, Cyp51A and Cyp51B. Overexpression and mutation of Cyp51A is a major cause of triazole resistance in A. fumigatus. The role of Cyp51B in generating resistance is unclear. Here, we show that overexpression or mutation of cyp51B results in triazole resistance. We demonstrate that introduction of a G457S Cyp51B mutation identified in a resistant clinical isolate results in voriconazole resistance in a naive recipient strain. Our results indicate that mutations in cyp51B resulting in clinical resistance do exist and should be monitored.

INTRODUCTION

Aspergillus fumigatus is the most common invasive mold pathogen in humans. It can cause a wide range of diseases in humans, with high mortality rates in immunocompromised patients (1, 2). The first line of treatment for invasive A. fumigatus infections is triazole antifungals that inhibit Erg11/Cyp51 lanosterol demethylase activity, blocking ergosterol biosynthesis. However, triazole-resistant strains of A. fumigatus are increasingly encountered, leading to increased mortality (3). To date, the most common triazole resistance mechanisms in A. fumigatus are alterations in the erg11A/cyp51A gene or promoter followed by overexpression of efflux pumps and mutations in hmg1, encoding 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase (3).

The A. fumigatus genome contains two cyp51 genes, cyp51A and cyp51B, showing 60% identity (4). Neither gene is essential, but deletion of both is lethal, suggesting they act in a compensatory fashion (5, 6). Deletion of cyp51A alone increases triazole susceptibility, whereas deletion of cyp51B has only a minor effect (6, 7). Cyp51A and Cyp51B green fluorescent protein (GFP)-tagged proteins localize primarily to the endoplasmic reticulum (ER), and their level of expression is upregulated after treatment with voriconazole (VRZ) (6). Expression of Cyp51A is strongly increased after deletion of Cyp51B, and vice versa, suggesting a potential compensatory response (6).

The involvement of Cyp51A in clinical triazole resistance has been well established and includes point mutations in codons G54, L98, Y121, T289, G138, and M220 and duplications in the promoter region (TR34 and TR46) that lead to overexpression (3, 8).

In contrast, the involvement of Cyp51B in clinical triazole resistance has not been fully revealed. cyp51B gene overexpression has been observed in two clinical triazole-resistant A. fumigatus isolates without cyp51A mutations, but a direct link between cyp51B overexpression and resistance was not made (9). Overexpression of cyp51B in a laboratory strain of A. fumigatus using a strong constitutive promoter resulted in moderately increased VRZ resistance (10). Recently, a clinical triazole-resistant A. fumigatus strain containing mutations in both Cyp51B (G457S) and Hmg1 (F390L) was described (11). However, the relative contribution of the cyp51B mutation toward resistance remained unclear.

Our goals, therefore, were to clearly define the contributions of A. fumigatus cyp51B overexpression and mutation toward triazole resistance. To achieve this, we expressed cyp51B under the inducible promoter xylP and under the cyp51A wild type (WT) and tandem repeat (TR)-containing promoter, in both WT and cyp51A null strains, and also introduced point mutations predicted to confer resistance. Our findings suggest that both cyp51B overexpression and mutation can lead to triazole resistance, indicating that this gene may have a more important role than previously realized in generating clinical triazole resistance. We show that the Cyp51B G457S mutation found in a clinical isolate confers voriconazole resistance when introduced into a susceptible strain, indicating a role for cyp51B as a generator of clinical resistance.

RESULTS

A. fumigatuscyp51A is more important than cyp51B in conferring triazole resistance.

Previous work has demonstrated that deletion of cyp51A alone increases triazole susceptibility, whereas deletion of cyp51B has only a minor effect (6, 7). We repeated, verified, and extended these findings. Deletion of cyp51A resulted in 4- to 16-fold increased sensitivity to the three triazoles tested, while deletion of cyp51B only slightly increased sensitivity (2-fold) to VRZ (Table 1). To quantify the effect of cyp51A expression levels on triazole susceptibility, we constructed two additional mutant strains with the tetOFF promoter regulating cyp51A expression in the WT (B+/A_tetOFF) or Δcyp51B background (B−/A_tetOFF) (Table 1). Strain B−/A_tetOFF under doxycycline (Dox) at 25 ng/ml and higher concentrations was inviable, indicating, as expected, that loss of both cyp51A and cyp51B was lethal in A. fumigatus. In both B+/A_tetOFF and B−/A_tetOFF strains, increasing Dox concentrations resulted in growing sensitivity to VRZ and itraconazole (ITZ), consistent with a progressive downregulation of the gene. Interestingly, strain B−/A_tetOFF was more sensitive than the B+/A_tetOFF strain under the same increasing Dox concentrations, indicating that cyp51B is important when cyp51A expression is compromised.

TABLE 1.

MIC values of voriconazole and itraconazole in WT and mutant strains with various doxycycline concentrations

Doxycycline concn (ng/ml)	MIC (mg/liter)
	VRZ in strain:				ITZ in strain:
	WT	Δcyp51B	B+/A_tetOFF	B−/A_tetOFF	WT	Δcyp51B	B+/A_tetOFF	B−/A_tetOFF
200	0.5	0.5	0.0625	Inviable	0.5	0.5	0.125	Inviable
100	0.5	0.5	0.0625	Inviable	0.5	0.5	0.25	Inviable
50	0.5	0.5	0.0625	Inviable	0.5	0.5	0.25	Inviable
25	0.5	0.5	0.0625	Inviable	0.5	0.5	0.25	Inviable
12.5	0.5	0.5	0.125	0.0625	0.5	0.5	0.25	0.0625
6.25	0.5	0.5	0.25	0.0625	0.5	0.5	0.25	0.0625
3.13	0.5	0.5	0.5	0.125	0.5	0.5	0.5	0.125
1.56	0.5	0.5	0.5	0.5	0.5	0.5	0.5	0.25
0.78	0.5	0.5	0.5	0.5	0.5	0.5	0.5	0.25
0.39	0.5	0.5	0.5	0.5	0.5	0.5	0.5	0.25
0	0.5	0.5	0.5	0.5	0.5	0.5	0.5	0.5

A. fumigatuscyp51B mRNA expression is triazole induced, but cyp51B is expressed at lower levels than cyp51A.

Previous studies have shown that A. fumigatus cyp51B and cyp51A can be induced by triazoles (6, 9). Because gene expression may vary among isolates, we used quantitative PCR (qPCR) to assess cyp51A and cyp51B expression in eight clinical strains, including ΔKU80 and Af293 laboratory strains, three triazole-susceptible strains (SCS1, SCS2, and SCS3), and three resistant strains (RCS1 [F46Y/M172K/E427K Cyp51A] and RCS-TR1 and RCS-TR2, both with TR34/L98H Cyp51A mutations) (see Tables S1 and S2 in the supplemental material). Mycelium was grown for 20 h and exposed to 0.5 MIC VRZ for 4 h, and total RNA was prepared for qPCR. The results show that VRZ induced the expression of cyp51A and cyp51B in three of the eight strains (ΔKU80, Af293, and SCS1) between 2- and 4-fold, indicating that induction is apparently strain specific at this time point and drug concentration (Fig. 1A and B).

FIG 1.

Expression of cyp51A and cyp51B in clinical isolates of A. fumigatus. Fold expression of cyp51A compared to that of cyp51B calculated as 2^−ΔΔCT. SCS, susceptible clinical strain; RCS, resistant clinical strain. RCS TR1 and RCS TR2 possess a Cyp51A TR34/L98H mutation, RCS1 possesses a Cyp51A F46Y/M172K/E427K mutation. (A) Fold induced expression of cyp51A after VRZ treatment (treated versus untreated). (B) Fold induced expression of cyp51B after VRZ treatment (treated versus untreated). (C) Fold basal expression of cyp51A compared to that of cyp51B in VRZ-untreated samples. (D) Fold induced expression of cyp51A compared to that of cyp51B in VRZ-treated samples. *, P < 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P < 0.0001.

Previous studies had not directly compared the basal and triazole-induced expression levels of cyp51A and cyp51B. Here, we show that cyp51A is expressed at higher basal and triazole-induced mRNA levels than cyp51B in all eight strains (Fig. 1C and D). Especially, high expression of cyp51A compared to that of cyp51B was observed in two resistant clinical strains (RCS-TR1 and RCS-TR2) resulting from a known Cyp51A TR34/L98H mutation that leads to overexpression.

Overexpression of A. fumigatus cyp51B leads to decreased triazole susceptibility.

Based on our results showing higher expression of cyp51A than of cyp51B, we hypothesized that increased expression of cyp51B may lead to triazole resistance. We therefore generated strains in which the cyp51B promoter was replaced (i) by the strong inducible promoter xylP in the ΔKU80 strain (generating strain A+/XylpB) and the Δcyp51A strain (generating strain A−/XylpB), (ii) by the cyp51Ap promoter (Ap) in the ΔKU80 strain (generating strain A+/ApB) and the Δcyp51A strain (generating strain A−/ApB), and (iii) by the cyp51A-TRp tandem repeat promoter (ATRp), in the ΔKU80 (generating strain A+/ATRpB) and Δcyp51A (generating strain A−/ATRpB) strains. Strain susceptibility to the triazoles VRZ, ITZ, and posaconazole (POS) was tested by broth microdilution (Table 2) and point inoculation on yeast extract-dextrose-MgSO₄ (YAG) agar plates, which we found to be more sensitive to minor changes in triazole susceptibility (Fig. 2). Using both methods, decreased triazole susceptibility was seen in both ΔKU80 and Δcyp51A backgrounds containing cyp51B under xylP (plus xylose) and cyp51A-TRp (Table 2 and Fig. 2). Moderately decreased triazole susceptibility was also seen by point inoculation in the ΔKU80 strain containing cyp51B under cyp51Ap (strains A+/ApB and A−/ApB) (Fig. 2A) and by broth microdilution in strain A−/ApB compared to that in the Δcyp51A strain (Table 2).

TABLE 2.

MIC values of voriconazole, itraconazole, and posaconazole in strains overexpressing cyp51B

Strain	MIC (mg/liter)
	VRZ	ITZ	POS
ΔKU80	1	0.5	0.25
A+/ApB	1	0.5	0.25
A+/ATRpBL113H	4	>16	1
A+/XylpB (+xylose)	4	>16	1
Δcyp51A	0.25	0.06	0.016
A−/ApB	1	0.25	0.125
A−/ATRpBL113H	4	2	0.5
A−/XylpB (+xylose)	4	1	0.25
Δcyp51B	0.5	0.5	0.25

FIG 2.

Triazole susceptibility of A. fumigatus strains overexpressing cyp51B. Droplet assay on RPMI-MOPS agar plates with increasing concentrations of VRZ, ITZ, and POS without (A and B) and with (B) xylose added to the medium. Each droplet column (from left to right) contained 10⁴, 10³, 10², or 10 spores per 10-μl droplet. A+/ApB, cyp51A-positive strain expressing cyp51B under cyp51A promoter; A−/ApB, cyp51A-negative strain expressing cyp51B under cyp51A promoter; A+/A_TRpB, cyp51A-positive strain expressing Cyp51B L113H under cyp51A TR34 promoter; A−/A_TRpB, cyp51A-negative strain expressing Cyp51B L113H under cyp51A TR34 promoter; A+/XylpB, cyp51A-positive strain expressing cyp51B under xylose promoter; A−/XylpB, cyp51A-negative strain expressing cyp51B under xylose promoter.

Point mutations in A. fumigatus cyp51B lead to decreased triazole susceptibility.

Point mutations in cyp51A are known to confer triazole resistance. However, whether this mechanism is also true for cyp51B is not clear. Therefore, we generated strains containing cyp51B mutations homologous to known triazole resistance mutations in Cyp51A. Mutations included (i) G69E (homologous to Cyp51A G54E), (ii) G153C (homologous to Cyp51A G138C), and (iii) G457S, identified in a clinical resistant A. fumigatus isolate (12) and homologous to Cyp51A G448S (see Fig. S10). Strain susceptibility to the triazoles VRZ, ITZ, and POS was tested by broth microdilution (Table 3) and point inoculation on YAG agar plates (Fig. 3). By broth microdilution, VRZ resistance was seen with Cyp51B G457S (4-fold increase in MIC versus that in the ΔKU80 control), while moderately decreased ITZ and POS susceptibilities (2-fold increase in MICs versus that in the ΔKU80 control) were seen with Cyp51B G153C. Moderately decreased POS susceptibility (2-fold increase in MIC versus that in the ΔKU80 control) was also seen with Cyp51B G69E (Table 3). By point inoculation, strongly decreased VRZ susceptibility with Cyp51B G457S and decreased ITZ and POS susceptibilities with Cyp51B G69E and Cyp51B G153C were seen (Fig. 3). By point inoculation, modestly increased VRZ susceptibility was seen with Cyp51B G69E and Cyp51B G153C, and modestly increased ITZ and POS susceptibilities were seen with Cyp51B G457S. The control Cyp51B WT-hph strain, in which a normal version of cyp51B had been introduced into the ΔKU80 strain, showed the same susceptibility pattern as the WT ΔKU80 strains. Strain growth in the absence of drug was not compromised by the mutations, as seen in the “no drug” control plate (Fig. 3).

TABLE 3.

MIC values of voriconazole, itraconazole, and posaconazole in strains with point mutations in cyp51B

Strain	MIC (mg/liter)
	VRZ	ITZ	POS
ΔKU80	0.5	0.25	0.25
Cyp51B G69E	0.5	0.25–0.5	0.5
Cyp51B G153C	0.5	0.5	0.5
Cyp51B G457S	2	0.25	0.25
Cyp51B WT-hph	0.5	0.25	0.25

FIG 3.

Triazole susceptibility of A. fumigatus strains with cyp51B mutations. Droplet assay on RPMI-MOPS agar plates containing increasing concentrations of VRZ, ITZ, and POS. Each droplet column (from left to right) contained 10⁴, 10³, 10², or 10 spores per 10- μl droplet.

DISCUSSION

Fungal enzymes of the Cyp51/ERG11 family mediate lanosterol demethylation, a crucial step in the biosynthesis of the fungal-specific sterol ergosterol. They have been intensively studied, as they are the target of azole antifungals, the main drug group used to treat fungal infections (13). However, due to their extensive medical use in the last 30 years and their massive application to treat mold infections in agricultural crops, resistance mutations in cyp51/ERG11 have emerged and are rapidly compromising treatment efficacy (8). Filamentous Pezizomycotina fungi such as the aspergilli are unique in that their genome contains 2 to 3 cyp51 copies (4, 14, 15). In A. fumigatus, which contains two cyp51 genes (cyp51A and cyp51B), cyp51A appears to be more dominant and important in generating azole resistance, while cyp51B has been considered more redundant and less involved in resistance (6, 7, 9, 10). The aim of this study was to better understand the role of cyp51B in generating triazole resistance in A. fumigatus.

In overview, our findings confirm that A. fumigatus cyp51B is less dominant than cyp51A in generating triazole resistance, as we show that deletion of cyp51A results in greater triazole susceptibility than deletion of cyp51B. Interestingly, the relevance of cyp51B is higher when cyp51A expression becomes compromised (Table 1). Therefore, it is tempting to speculate that the presence of cyp51B in A. fumigatus may have facilitated the development of mutations in cyp51A, serving as a backup system to maintain some level of resistance if Cyp51A function becomes compromised after mutation. Additionally, we provide several lines of evidence suggesting that this is primarily because cyp51B is driven by a weaker promoter: First, in multiple clinical strains, cyp51B is expressed at lower levels than cyp51A. Second, overexpression of cyp51B under the inducible xyl promoter or TR-cyp51A promoter generates substantial triazole resistance. Third, expression of cyp51B under the native cyp51A promoter in a strain lacking cyp51A (strain A−/ApB) (Table 2 and Fig. 2) results in greater resistance than a strain containing only cyp51A (strain Δcyp51B) (Table 2 and Fig. 2).

If, as we have shown, overexpression of cyp51B can lead to triazole resistance, why has it not been found in clinical isolates? We believe that a very real possibility is that it has been overlooked. An additional explanation, although not exclusive, is that mutations in the cyp51B promoter that lead to overexpression may be difficult to generate. cyp51A overexpression occurs by tandem-repeat duplication of the two serum response element (SRE) promoter binding sites to generate four SRE sites, thereby increasing binding by the positive transcriptional regulator SrbA (16). The cyp51B promoter contains a single SRE site (see Fig. S11 in the supplemental material) that could be duplicated under selection, but it is unclear if two SRE sites will generate sufficiently high levels of cyp51B expression to generate resistance. Future research should aim to investigate this possibility in the laboratory.

To explore the possibility that point mutations in A. fumigatus cyp51B can confer triazole resistance, we introduced mutations G69E and G153C, homologous to the Cyp51A mutations G54E and G138C, respectively. The G54E Cyp51A mutation is situated in the substrate access channel and appears to interact directly with the triazole drug. It leads to ITZ and POS resistance but typically does not affect VRZ susceptibility (3). Interestingly, the G69E Cyp51B also resulted in up to 2-fold increase in the MICs to ITZ and POS but not to VRZ. Point inoculation analysis, which we find is more sensitive than the standard MIC test, showed increased VRZ susceptibility in the Cyp51B G69E mutant.

The G138C Cyp51A mutation is located in the substrate access channel near the heme cofactor and leads to VRZ, POS, and ITZ resistance (10, 17). The analogous Cyp51B G153C mutation resulted in an up to 2-fold increase in the MICs to ITZ and POS but not to VRZ. Taken together, our results show that cyp51B mutations mimic their analogous cyp51A mutations with respect to their triazole resistance profile, but, probably because of reduced expression of cyp51B, they achieve a lower level of resistance. It will be interesting to investigate if combinations of tandem repeats (TRs) in the promoter region with point mutations in the open reading frame (ORF) can confer higher levels of resistance, as is the case with the TR34/L98H and the TR46/Y121F/T289A mutations in cyp51A.

The most important finding described here is that a cyp51B point mutation occurring in a clinical resistant isolate is responsible for triazole resistance. The Cyp51B G457S mutation was originally described in the pan-azole-resistant strain CM9460, which also contains an F390L mutation in Hmg1. The glycine in position 457 in Cyp51B is at the same position as glycine 448 in Cyp51A, which when mutated to serine, leads to VRZ resistance and weak ITZ and POS resistance, apparently by altering the conformation of the heme-binding site (11, 12). Here, we show that reconstitution of the Cyp51B G457S mutation in a triazole-sensitive strain of A. fumigatus results in resistance to VRZ (MIC increased 4-fold to 2 μg/ml) but not to ITZ or POS. In contrast, strain CM9460 showed MICs of >8 μg/ml for VRZ, ITZ, and POS (11), which suggests that either its resistance is mainly driven by the mutation in hmg1 or simultaneous presence of both mutations has a synergistic effect on triazole resistance, as has been suggested for hmg1 and cyp51A. This should be confirmed by reconstitution of mutated cyp51B and hmg1 alone and in combination in a triazole-susceptible strain. Taking into consideration that the cyp51B mutations described here have a smaller role in generating triazole resistance than parallel cyp51A mutations, we predict that they will be important as secondary drivers alongside stronger mutations in cyp51A and hmg1.

In summary, mutations in cyp51B can lead to triazole resistance in A. fumigatus, and we suggest this should be considered when analyzing resistant clinical isolates.

MATERIALS AND METHODS

Media and strains.

Strains were grown on YAG agar plates (0.5% yeast extract, 1% dextrose, 0.01 M MgSO₄, trace elements solution, vitamin mix, 1.5% agar) for 48 to 72 h. MIC experiments were performed in RPMI-MOPS broth (10% 10× RPMI medium, 3.45% morpholinepropanesulfonic acid [MOPS]), and droplet assays were performed on RPMI-MOPS agar plates. Transformation of A. fumigatus spores was made on YPGS agar plates (2% yeast extract, 0.5% peptone, 2% d-glucose, 1 M sucrose, 1.5% agar for plates or 0.7% for top, pH 6) with 200 μg/ml hygromycin B for hygromycin B selection or on AMM sucrose agar plates (1× salts solution, 1% dextrose, 1 M sucrose, 0.012 M KPO₄, pH 6.8, 0.1% trace elements solution, 1.5% agar for plates or 0.7% for top, pH 6) with 0.1 μg/ml pyrithiamine for pyrithiamine selection.

This study included the wild-type strains Af293 and ΔKU80, six patient isolates (see Table S1 in the supplemental material), cyp51A and cyp51B single knockout strains constructed from the ΔKU80 strain, cyp51B promoter replacement strains constructed from either ΔKU80 or Δcyp51A strains, and cyp51B point mutation strains constructed from the ΔKU80 strain. A detailed description of the construction and verification of these strains is provided in the Tables S1, S3, and S4 and Fig. S1 to S9.

Antifungal susceptibility testing.

MIC was determined by CLSI M38-A2 broth microdilution methodology. Briefly, stock solutions of either voriconazole (VRZ), itraconazole (ITZ), or posaconazole (POS) (Merck Sigma) were diluted in RPMI-MOPS and loaded into 96-well plates; spores were diluted to 5 × 10⁴ spores/ml and were loaded into the wells. Plates were incubated at 37°C for 48 h, and then the lowest concentration of triazole completely inhibiting fungal growth (observed by inverted light microscope) was set as the MIC. The CLSI epidemiological cutoff values (ECVs) used were >1 mg/liter for itraconazole, >1 mg/liter for voriconazole, and >0.25 mg/liter for posaconazole (18). Droplet susceptibility testing was performed by inoculation of 10⁴, 10³, 10², or 10 spores in 10 μl distilled water on the surface of RPMI-MOPS agar plates containing different concentrations of antifungal and incubation for 48 h at 37°C.

Determination of gene expression by qPCR.

Strains were grown on YAG agar plates for 72 h, and 6 × 10⁷ spores were collected and inoculated in 150 ml YAG broth in 250-ml flasks. Flasks (6 per strain) were incubated under shaking for 20 h at 37°C, and then 0.5 MIC of VRZ (Table S2) was added to 75 ml for 4 h, and 75 ml was left untreated. Mycelium was collected, lyophilized, and crushed; RNA was extracted using the Qiagen RNeasy plant minikit. RNA concentration was assessed using a NanoDrop, and then equal amounts of RNA from each sample were converted to cDNA using the Verso cDNA synthesis kit. Equal amounts of cDNA (based on RNA amounts) from each sample were loaded into Applied Biosystems MicroAmp optical 96-well plates with Applied Biosystems Fast SYBR green master mix and primer for either β-tubulin (housekeeping control gene), cyp51A, or cyp51B. Comparative threshold cycle (2^−ΔΔCT) analysis was performed. Statistical analysis was performed with one-way analysis of variance (ANOVA).