hapE and hmg1 Mutations Are Drivers of cyp51A -Independent Pan-Triazole Resistance in an Aspergillus fumigatus Clinical Isolate

Aspergillus fumigatus is a ubiquitous environmental mold that can cause severe disease in immunocompromised patients and chronic disease in individuals with underlying lung conditions. Triazoles are the most widely used class of antifungal drugs to treat A. fumigatus infections, but their use in the clinic is threatened by the emergence of triazole-resistant isolates worldwide, reinforcing the need for a better understanding of resistance mechanisms. The predominant mechanisms of A. fumigatus triazole resistance involve mutations affecting the promoter region or coding sequence of the target enzyme of the triazoles, Cyp51A. However, triazole-resistant isolates without cyp51A -asso-ciated mutations are frequently identi ﬁ ed. In this study, we investigate a pan-triazole-resistant clinical isolate, DI15-105, that simultaneously carries the mutations hapE P88L and hmg1 F262del , with no mutations in cyp51A . Using a Cas9-mediated gene-editing system, hapE P88L and hmg1 F262del mutations were reverted in DI15-105. Here, we show that the combination of these mutations accounts for pan-triazole resistance in DI15-105. To our knowledge, DI15-105 is the ﬁ rst clinical isolate reported to simultaneously carry mutations in hapE and hmg1 and only the second with the hapE P88L mutation. IMPORTANCE Triazole resistance is an important cause of treatment failure and high mortality rates for A. fumigatus human infections. Although Cyp51A-associated mutations are frequently identi ﬁ ed as the cause of A. fumigatus triazole resistance, they do not explain the resistance phenotypes for several isolates. In this study, we demonstrate that hapE and hmg1 mutations additively contribute to pan-triazole resistance in an A. fumigatus clinical isolate lacking cyp51 -associated mutations. Our results exem-plify the importance of and the need for a better understanding of cyp51A -independ-ent triazole resistance mechanisms.

The predominant mechanism of triazole resistance in A. fumigatus is genetic mutation leading to changes in the amino acid sequence and/or expression levels of one of the two target enzymes, Cyp51A (19,20). Tandem repeats (TRs) in the promoter of the cyp51A gene coupled with single point mutations in the coding sequence, such as TR 34 / L98H or TR 46 /Y121F/T289A, have been widely reported in clinical and environmental isolates. Such TRs in the promoter lead to upregulation of cyp51A expression, which in turn increases the amount of triazole necessary for enzyme inhibition (21,22). Moreover, amino acid substitutions in hot spot residues, such as G54 and M220, as well as those occurring with TRs in the promoter, are thought to cause conformational changes that reduce the binding affinity of the triazole drugs with the target enzyme (19,20,23,24).
While cyp51A-associated mutations are the most common and best-characterized mechanism of triazole drug resistance, non-cyp51A triazole-resistant clinical isolates have been increasingly reported, with prevalence rates varying from 15 to 60% depending on geographic region (25)(26)(27)(28). However, cyp51A-indepent triazole drug resistance remains poorly understood and underinvestigated. One of the potential causes of triazole drug resistance is the overexpression of ATP binding cassette (ABC) and major facilitator superfamily (MFS) multidrug transporters (29)(30)(31), which is often observed among triazole-resistant A. fumigatus isolates. Overexpression of these transporters is thought to cause a decrease in intracellular drug concentrations, therefore decreasing the amount of drug available to inhibit sterol demethylases (29)(30)(31).
Mutations in hmg1, which encodes the 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase Hmg1, have also recently emerged as a genetic determinant of triazole drug resistance in A. fumigatus (26). Hmg1 initiates ergosterol biosynthesis by catalyzing the reduction of HMG-CoA into mevalonate and has a conserved motif known as the sterol-sensing domain (SSD), which is predicted to participate in the regulation of HMG-CoA reductase activity (32)(33)(34). Although the clinical prevalence of hmg1 mutations has not been assessed in a systematic manner, such mutations have been identified in over 150 isolates thus far reported in the literature, many of which are resistant to triazole drugs and lack cyp51A-associated mutations (25)(26)(27)(28)(35)(36)(37)(38)(39)(40)(41)(42)(43)(44). The most frequent Hmg1 amino acid substitutions associated with triazole drug resistance occur in residues putatively located in the Hmg1 SSD, such as S269, S305, G307, and I412, with several studies showing that some of these mutations, namely, F262del, S305P, I412S, S269F, and E306K, impart decreased susceptibility to multiple triazoles when genetically introduced into a susceptible laboratory strain (25,26,36,40). The mechanism by which hmg1 mutations alter triazole resistance remains to be defined; however, it is speculated that amino acid substitutions in the SSD impair its ability to sense sterols and signal for Hmg1 degradation (25,26,28).
Another recently identified mechanism of triazole drug resistance involves mutations in the hapE gene, which encodes one of the three subunits of the CCAAT box binding complex (CBC), a transcription factor that has been shown to repress the expression of several ergosterol biosynthesis genes, including cyp51A (45,46). It has been suggested that the hapE P88L mutation decreases the binding affinity of the CBC for the promoter region of cyp51A (47), impairing its ability to repress gene expression, which culminates in overexpression of cyp51A. Thus far, hapE mutations seem to be rare occurrences, having been reported for only 3 isolates (26,46,48).
In this study, we investigated a pan-triazole-resistant clinical isolate, DI15-105, that simultaneously carries the mutations hapE P88L and hmg1 F262del and lacks mutations in cyp51A. Our goal was to determine the impact of these mutations individually and in combination on the triazole resistance phenotype of DI15-105. Using a Cas9-mediated gene-editing system, hapE P88L and hmg1 F262del mutations were reverted in DI15-105. Our results demonstrated that, together, these mutations account for the high levels of pan-triazole resistance in DI15-105.

RESULTS
Correction of the hapE P88L and hmg1 F262del mutations recovers DI15-105 susceptibility to triazole drugs. The hapE P88L and hmg1 F262del mutations have been individually observed in triazole-resistant clinical isolates and independently shown to influence triazole susceptibility (25,26,46). Using in vitro-assembled Cas9 ribonucleoproteins (RNPs), allele swaps were performed and successfully replaced the mutated versions of hapE and hmg1 (previously performed [26]) with their respective wild-type (WT) versions (derived from AF293) in the clinical isolate DI15-105, as confirmed by PCR screening (see Fig. S1 and S2 in the supplemental material) and Sanger sequencing (data not shown). Using gradient diffusion test strips ( Fig. 1), triazole susceptibility was determined and revealed that correction of either hapE or hmg1 gene sequences decreased DI15-105 MICs by 2fold and 5-fold, respectively. MIC values decreased even further, $10-fold, when hapE and hmg1 gene sequences were both corrected.
Correction of the hapE P88L and hmg1 F262del mutations in DI15-105 reduces cyp51A expression levels. It has been suggested that the hapE P88L mutation reduces the affinity of CBC for the promoter region of cyp51A, thus impairing its ability to repress gene expression. This ultimately leads to an increase in cyp51A expression, resulting in increased MICs for triazole drugs. It is not clear by what mechanism hmg1 mutations, particularly hmg1 F262del , increase triazole resistance, and previous studies have not reported changes in cyp51A expression. Since we observed an additive effect of these two mutations contributing to triazole resistance, we assessed expression levels of cyp51A upon correction of hapE P88L and hmg1 F262del mutations. As expected, correction of the hapE mutation in both DI15-105_hapE WT-PhleoR /hmg1 F262del and DI15-105_hapE WT-PhleoR /hmg1 WT-HygR significantly decreased expression of cyp51A after 24 h (;2.8-fold and ;2.7-fold, respectively) and 48 h (;1.5-fold and ;2-fold, respectively) of fungal growth (Fig. 3). Correction of hmg1 also caused a less prominent but statistically significant decrease in cyp51A expression at 24 h (;1.4-fold) and 48 h (;1.3-fold) of fungal growth.
Correction of the hapE P88L mutation improves DI15-105 radial growth and induces hypersensitivity to oxidative stress. In some studies, the hapE P88L mutation has been associated with decreased fungal growth (45,47). Therefore, we analyzed colony macromorphology (Fig. 4A) and radial growth (Fig. 4B) of the DI15-105 isolate and derived strains. Macromorphological analysis revealed that the strain with a hapE single correction (DI15-105_hapE WT-PhleoR /hmg1 F262del ) presented significantly enhanced radial growth, which was not observed in the double correction strain DI15-105_hapE WT-PhleoR / hmg1 WT-HygR . We also observed that the hapE correction strains DI15-105_hapE WT-PhleoR / hmg1 F262del and DI15-105_hapE WT-PhleoR /hmg1 WT-HygR presented a darker green color than hapE P88L strains, suggesting greater conidium production; however, quantification of conidia did not show significant differences in the amounts of conidia per colony area among the isolates (Fig. S3). In order to evaluate the impact of hmg1 and hapE mutations on the DI15-105 stress response, strains were grown in agar medium supplemented with several stressors, namely, H 2 O 2 (oxidative stress) (Fig. 4C), iron (depletion and excess amounts) (see Fig. S4A), calcofluor white and Congo red (cell wall stress) (see Fig. S4B), and sorbitol (osmotic stress) (see Fig. S4B). Only correction of the hapE P88L mutation in DI15-105_hapE WT-PhleoR /hmg1 F262del rendered the isolate hypersensitive to oxidative stress, which once again was not observed when the hmg1 mutation was also corrected ( Fig. 4; also see Fig. S5).

DISCUSSION
The emergence of triazole drug resistance represents an important challenge in the therapy of A. fumigatus infections (50). Genetic alterations in the cyp51A gene coding sequence or promoter region are the most frequently identified and best-characterized   (26,46). Here, our goal was to determine the impact of these mutations individually and in combination on the triazole resistance phenotype of DI15-105. The hmg1 F262del mutation is predicted to alter a residue in the SSD of Hmg1, a conserved motif shown to be essential for the negative regulation of sterol biosynthesis in other eukaryotes, including Schizosaccharomyces pombe (32,33,52). We previously observed that restoration of the hmg1 WT sequence in DI15-105 restored triazole susceptibility to all mold-active triazoles (26). The same hmg1 F262del mutation introduced in a susceptible A. fumigatus laboratory strain induced increases of at least 4-fold in the MICs for all triazoles, accumulation of ergosterol precursors, and significant increases in total cellular ergosterol levels (26). In a similar way, amino acid substitutions in other residues of the Hmg1 SSD have also been identified, and they are often associated with triazole-resistant isolates (25-27, 35, 36). Some of these mutations, namely, S305P, I412S, S269F, and E306K, have been shown to impart decreased susceptibility to multiple triazoles when genetically inserted in a triazole-susceptible laboratory strain (25,26,36). It is important to note that many of the hmg1 mutations so far reported in the literature occur in isolates that also carry cyp51A mutations that are known to drive triazole drug resistance (25)(26)(27)(28)36). However, it was demonstrated that, in some cases, the presence of the hmg1 mutation led to atypical and increased triazole MIC levels in the isolates (26,27). For instance, genetic reversion of the cyp51A mutations TR 34 /L98H, G138C and M263I to the WT sequence did not recover triazole susceptibility in different clinical isolates (26). Those isolates also carried the hmg1 mutations E105K, Y250H, and I412S, respectively; the latter was previously shown to contribute directly to triazole resistance (26). Additionally, in a clinical isolate carrying the cyp51A M220I amino acid substitution, reversion of the hmg1 S305P mutation reduced the VRC MIC while maintaining ITRA and POS MIC values (36). These data support the evidence that hmg1 mutations are important determinants of clinical triazole resistance. However, the mechanism by which Hmg1 SSD mutations lead to triazole resistance remains to be fully elucidated, and it is speculated that such mutations might affect the negative regulation of HMG-CoA reductase activity (25,26).
The hapE P88L mutation present in DI15-105 was previously reported in a single clinical isolate that did not also carry mutations in cyp51A (46). The hapE gene encodes one of the three subunits of the CBC, a highly conserved transcription factor that has been The cyp51A expression levels of hmg1 and hapE correction strains were assessed using RT-qPCR. RPMI 1640 medium was inoculated with A. fumigatus conidia at a concentration of 2 Â 10 4 conidia/mL and incubated at 35°C and 250 rpm for 24 h or 48 h. Mycelia were harvested and subjected to RNA extraction. cDNA was synthesized, and qPCR was performed using tubA, which encodes b-tubulin A, as the housekeeping gene for data normalization. Changes in gene expression among isolates were calculated using the 2 2DDCT method. Experiments were performed using four biological replicates. Statistical analysis was performed using one-way ANOVA with Tukey's multiplecomparison test. **, P , 0.01; ***, P , 0.001; ****, P , 0.0001. shown to repress expression of several ergosterol biosynthesis genes, such as cyp51A, cyp51B, erg7B, and erg13B (45). It has been demonstrated that the hapE P88L mutation decreases the binding affinity of the CBC for the promoter region of cyp51A, impairing its ability to repress gene expression, which culminates in overexpression of cyp51A and thus increased tolerance to triazoles (47). Our results show that the single correction of the hapE P88L mutation in DI15-105_hapE WT-PhleoR /hmg1 F262del leads to significant decreases in triazole MICs. Recently, a triazole-resistant clinical isolate without cyp51A mutations but carrying a HapE splicing site mutation (c.154-1G.A) was identified; it presented higher cyp51A expression levels than did an isogenic and triazole-susceptible isolate that had been collected previously from the same patient (48). However, the direct effect of this hapE splicing site mutation on triazole resistance has not yet been characterized. Interestingly, in both reports, hapE mutations in clinical isolates were thought to have been acquired in the patient after prolonged therapy with triazoles (47,48). To our knowledge, DI15-105 is only the second isolate that has been reported to carry the hapE P88L mutation (26). These examples reinforce the clinical importance of hapE mutations for triazole resistance. In our study, we observed that the hmg1 F262del mutation has a major impact on the DI15-105 triazole resistance phenotype, since the reversion of this mutation generated a greater decrease in triazole MICs than did the correction of the hapE P88L mutation, in both strip and broth microdilution assays. In addition, correction of both hmg1 F262del and hapE P88L mutations in DI15-105_hapE WT-PhleoR /hmg1 WT-HygR led to a greater reduction in triazole MICs, demonstrating the additive effects of the two mutations on triazole resistance. In order to elucidate how the two mutations enhanced triazole resistance, we evaluated whether the increased susceptibility of the correction strains to triazole drugs was a consequence of cyp51A expression downregulation. Reverse transcription-quantitative PCR (RT-qPCR) results showed a statistically significant but slight decrease in cyp51A expression in all correction strains, although the decrease in cyp51A expression levels was more prominent in the DI15-105_hapE WT-PhleoR /hmg1 F262del and DI15-105_hapE WT-PhleoR /hmg1 WT-HygR strains, in which the hapE mutation was corrected.
Given that the single correction of the hapE P88L mutation did not reduce triazole MICs to the extent observed with the reconstitution of both hapE and hmg1 mutations, we think that cyp51A expression is not the only mechanism operating in DI15-105. Further investigation is needed to demonstrate how the presence of the hmg1 F262del and hapE P88L mutations together leads to increased MIC values in DI15-105. In order to identify other ways in which hmg1 F262del and hapE P88L mutations affect A. fumigatus, we analyzed fungal growth and responses to oxidative, iron, osmotic, and cell wall stress in DI15-105 and derived strains. While no significant changes were observed in the DI15-105_hapE P88L /hmg1 WT-HygR and DI15-105_hapE WT-PhleoR /hmg1 WT-HygR strains, surprisingly a slight improvement in radial growth was detected in the strain in which only the hapE P88L mutation was corrected. Although it was not observed in DI15-105_hapE WT-PhleoR /hmg1 WT-HygR , this recovery of radial growth observed after single correction of hapE is consistent with previous reports that showed that the hapE P88L mutation, as well as the gene deletion of other subunits of the CBC, generate impairment of fungal growth (45,47). Moreover, while no unique phenotypes were displayed when strains were grown under iron, osmotic, or cell wall stress conditions, single correction of the hapE P88L mutation in DI15-105_hapE WT-PhleoR /hmg1 F262del resulted in hypersensitivity to H 2 O 2 . The fact that these phenotypes appear only in the strain in which single correction of hapE was performed is intriguing and suggests that, in DI15-105_hapE WT-PhleoR /hmg1 F262del , recovery of hapE ability to regulate gene expression favors fungal growth under basal conditions while also disturbing the strain's ability to respond to oxidative stress when the hmg1 F262del mutation is present. Further investigation is required for clarification of this phenomenon.
In summary, we report for the first time that, in a clinical isolate with no cyp51Aassociated mutations, hmg1 F262del and hapE P88L mutations additively contribute to pantriazole resistance. Because cyp51A expression levels do not seem to entirely explain the phenotypes observed in this study, the mechanism by which the two mutations lead to A. fumigatus triazole resistance remains unclear. This study is another example of the clinical importance of non-cyp51A-associated triazole resistance and reinforces the need for a better understanding of the mechanisms that drive triazole resistance in A. fumigatus.

MATERIALS AND METHODS
Isolates and growth conditions. The pan-triazole-resistant clinical isolate DI15-105 was kindly donated by Nathan P. Wiederhold at the Fungus Testing Laboratory at the University of Texas Health Science Center at San Antonio. In a previous study, whole-genome sequencing revealed that this isolate presented no mutations in cyp51A but carried the hapE P88L and hmg1 F262del mutations (26). In the same study, a DI15-105-derived strain (DI15-105_ hapE P88L /hmg1 WT-HygR ) in which the hmg1 F262del mutation was reconstituted with a WT gene sequence was created, and it is also used here.
Fresh conidia from each isolate were obtained after harvesting of 3-day-old glucose minimal medium (GMM) agar cultures at 37°C. The concentrations of conidia in water stocks were then determined using a hemocytometer, and suspensions were kept at 4°C or used to prepare 50% glycerol stocks. Radial growth was determined by point inoculation of GMM plates with a 5-mL drop of a conidial suspension of 2 Â 10 6 conidia/mL, and colony diameters were measured after incubation at 37°C for 96 h. Statistical analysis was performed using one-way analysis of variance (ANOVA) followed by Tukey's multiple-comparison test with GraphPad Prism 9 software.
Gene-editing strategy and A. fumigatus transformation mediated by CRISPR/Cas9. In order to correct the hapE P88L and hmg1 F262del mutations in the clinical isolate DI15-105, allele swaps mediated by a Cas9 system were performed, as described previously (26,53). For reference, all CRISPR RNA (crRNA) sequences and primers (Integrated DNA Technologies) used to build strains are described in Table S1 in the supplemental material, and a detailed scheme of the gene-editing strategy is depicted in Fig. S1 and hapE and hmg1 Mutations and Triazole Drug Resistance Microbiology Spectrum S2. Repair templates were constructed using WT hapE (Afu6g05300) and hmg1 (Afu2g03700) gene sequences obtained from AF293 genomic DNA. For hapE gene manipulation, the WT gene open reading frame (ORF) was cloned into the pAGRP plasmid (54), which contains a phleomycin resistance cassette (PhleoR), using the restriction enzymes NotI and PacI. After confirmation of successful insertion of hapE by diagnostic digestion and Sanger sequencing (data not shown), the plasmid (pAGRP-hapE) was used as a template for amplification of the repair template, with primers that incorporated 35-bp microhomology regions with the hapE ORF at both the 59 and 39 ends. The DI15-105_ hapE P88L /hmg1 WT-HygR strain (26) was created previously, as follows. For hmg1 gene manipulation, a two-component repair template that consisted of a split hmg1 allele and a hygromycin resistance marker (HygR) was used. A WT hmg1 gene sequence, including the ORF and around 50 bp upstream and 500 bp downstream, was PCR amplified using a 39 primer that inserted 80 bases with homology to the 39 end of the hygromycin B resistance gene ORF. The HygR cassette was PCR amplified from the pUCGH plasmid (55) using a 39 primer that introduced approximately 70 bases with homology to the downstream region of hmg1.
In order to replace the whole hmg1 or hapE native ORF with a sequence of interest, two separate guide RNAs (gRNA) were selected to generate double-stranded DNA breaks at the 59 and 39 regions of the target gene ORF. gRNA duplexes were built using equimolar concentrations of a gene-specific 59 or 39 crRNA (see Table S1) and a universal transactivating crRNA (tracrRNA), and Cas9 RNP complexes were then assembled in vitro by mixing both 59 and 39 gRNA duplexes with the Cas9 enzyme, according to the manufacturer's instructions. The transformation was carried out by mixing Cas9 RNPs with protoplasts generated from DI15-105 mycelia through enzymatic digestion and 2 to 5 mg of the desired repair template, as described in detail elsewhere (53). Transformation reactions were carried out on sorbitol minimal medium (SMM) agar plates (1.5% agar), and protoplasts were allowed to recover overnight at room temperature. For hmg1 transformations, SMM top agar (0.75% agar) containing 450 mg/mL hygromycin was added to the plates and incubated at 37°C for 3 to 5 days. For hapE transformations, SMM top agar with 375 mg/mL phleomycin was added to the plates and incubated until the next day at room temperature, followed by incubation at 30°C for 3 to 5 days. Single colonies were then subcultured in GMM supplemented with 150 mg/mL hygromycin or 125 mg/mL phleomycin, genomic DNA was extracted, and mutants were confirmed by multiple screening PCRs (see Fig. S1 and S2) and Sanger sequencing (data not shown).
The genetic manipulations performed in this study generated strain DI15-105_hapE WT-PhleoR /hmg1 F262del , in which the mutated hapE allele was replaced with the WT gene sequence, and strain DI15-105_hapE WT-PhleoR / hmg1 WT-HygR , in which both mutated hapE and hmg1 alleles were replaced with the respective WT gene sequences. The previously constructed DI15-105_hapE P88L /hmg1 WT-HygR strain (26) was used as a background to generate the double correction strain DI15-105_hapE WT-PhleoR /hmg1 WT-HygR .
Antifungal susceptibility testing. Antifungal susceptibility to all clinically available mold-active triazoles was determined using gradient diffusion test strips and broth microdilution assays. VRC, ISA, ITRA, and POS strips were applied in the center of RPMI 1640 agar plates that had been inoculated with 500 mL of a suspension of 2 Â 10 6 conidia/mL. The plates were incubated for 48 h at 35°C, and the MIC was defined where a zone of growth inhibition intercepted the strip reading scale. Triazole broth microdilution assays were performed according to CLSI M38-A2 methodology with modifications (49). Briefly, 2-fold serial dilutions of triazole drugs were distributed in U-bottomed 96-well plates, followed by a suspension of A. fumigatus conidia with a final concentration of 2 Â 10 4 conidia/mL. The final drug concentrations tested varied from 0.03125 to 32 mg/mL. The plates were incubated for 48 h at 35°C, and MICs were determined.
Determination of cyp51A expression levels. The cyp51A expression levels of DI15-105 and derived strains were assessed using RT-qPCR. Briefly, RPMI 1640 medium was inoculated with A. fumigatus conidia at a concentration of 2 Â 10 4 conidia/mL and incubated at 35°C and 250 rpm for 24 h or 48 h. Cultures supernatants were discarded, and mycelia were harvested, washed in sterile distilled water, placed in screw-cap tubes, and immediately frozen in liquid N 2 . RNA extraction was performed using the RiboPure-Yeast kit (Invitrogen) according to the manufacturer's instructions, followed by an extra purification step with isopropanol. RNA concentrations were determined using a NanoDrop spectrophotometer, and 0.5 mg of RNA was used for cDNA synthesis with the RevertAid first-strand cDNA synthesis kit (Thermo Fisher Scientific). qPCR was performed using the SsoAdvanced universal SYBR green supermix (Bio-Rad) with 1 mL of a 1:50 cDNA dilution as the template. The tubA gene, which encodes b-tubulin A, was selected as the housekeeping gene for data normalization, and changes in gene expression among isolates were calculated using the 2 2DDCT method with the average DC T of DI15-105 as the control. Primers are listed in Table S1 in the supplemental material. Experiments were performed using four biological replicates, each analyzed in technical duplicates. Statistical analysis was performed using oneway ANOVA followed by Tukey's multiple-comparison test with GraphPad Prism 9 software.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 2.9 MB.

ACKNOWLEDGMENTS
This work was supported by NIH grant R01 AI143197 to P.D.R. and J.R.F. A.C.O.S. performed experiments, interpreted results, and wrote the manuscript. W.G. performed experiments. N.P.W. provided critical input and feedback. J.M.R. interpreted results, provided critical feedback, and assisted in writing the manuscript. J.R.F. and P.D.R. interpreted results, provided critical feedback, assisted in writing the manuscript, and supervised the study.