Differential Functions of Individual Transcription Factor Binding Sites in the Tandem Repeats Found in Clinically Relevant cyp51A Promoters in Aspergillus fumigatus.

Aspergillus fumigatus is the major filamentous fungal pathogen in humans. The gold standard treatment of A. fumigatus is based on azole drug use, but the appearance of azole-resistant isolates is increasing at an alarming rate. The cyp51A gene encodes the enzymatic target of azole drugs, and azole-resistant alleles of cyp51A often have an unusual genetic structure containing a duplication of a 34- or 46-bp region in the promoter causing enhanced gene transcription. These tandem repeats are called TR34 and TR46 and produce duplicated binding sites for the SrbA and AtrR transcription factors. Using site-directed mutagenesis, we demonstrate that both the SrbA (sterol response element [SRE]) and AtrR binding sites (AtrR response element [ATRE]) are required for normal cyp51A gene expression. Loss of either the SRE or ATRE from the distal 34-bp repeat of the TR34 promoter (further 5' from the transcription start site) caused loss of expression of cyp51A and decreased voriconazole resistance. Surprisingly, loss of these same binding sites from the proximal 34- or 46-bp repeat led to increased cyp51A expression and voriconazole resistance. These data indicate that these duplicated regions in the cyp51A promoter function differently. Our findings suggest that the proximal 34- or 46-bp repeat in cyp51A recruits a corepressor that requires multiple factors to act while the distal repeat is free of this repression and provides the elevated cyp51A expression caused by these promoter duplications. IMPORTANCE Aspergillus fumigatus is the most common human filamentous fungal pathogen. Azole drugs are the current therapy of choice for A. fumigatus, but the prevalence of azole resistance is increasing. The main genetic alteration causing azole resistance is a change in the cyp51A gene, which encodes the target of these drugs. Azole-resistant cyp51A alleles routinely contain duplications in their promoter regions that cause increased gene transcription. Here, we demonstrate that clinical isolates containing a 34- or 46-bp duplication in the cyp51A promoter required the presence of the transcription factor-encoding atrR gene to exhibit elevated azole resistance. Eliminations of transcription factor binding sites in the cyp51A gene have differential actions on expression of the resulting mutant allele. These data dissect the molecular inputs to cyp51A transcription and reveal a complicated function of the promoter of this gene that is critical in azole resistance.

OBSERVATION

Aspergillus fumigatus is the most common cause of mold infections in humans (1). Azole drugs are currently the first-line therapy for aspergillosis. However, azole-resistant A. fumigatus clinical isolates are being found with increasing frequency and are associated with a significantly worse clinical outcome (2). Although multiple mechanisms contribute to azole resistance in A. fumigatus, the most commonly reported genetic changes associated with this phenotype are alterations in the gene encoding Cyp51A, the target enzyme of azole drugs (3). The most prevalent azole resistance allele is a compound mutation in cyp51A consisting of a 34-bp duplication in the promoter element (TR34) and a single amino acid replacement in the coding sequence (L98H) (4). Both of these mutations are required for the observed high-level azole resistance conferred by this compound allele (5).

Although it is well established that the TR34 cyp51A promoter drives increased expression of cyp51A mRNA compared to the wild-type version (5), we lack a detailed understanding of how this increased expression is achieved. Previous studies from our lab and others have demonstrated that several different transcription factors control transcription of cyp51A via the 34-bp region. First, the sterol-responsive SrbA regulator binds to an element in this 34-bp region called the sterol response element (SRE) and stimulates expression when sterols are limiting (6, 7). Second, the AtrR transcription factor binds to a second site within the 34-bp region, referred to as the AtrR response element (ATRE), to activate transcription (8, 9). Finally, two different negative transcriptional regulators repress cyp51A expression. The CCAAT-binding complex (CBC) and the iron-responsive transcription factor HapX (10) both reduce cyp51A expression: CBC binds within the 34-bp region, while HapX binds just 3′ to this segment (11, 12). The locations of these sites and their positions relative to the 34-bp region are shown in Fig. S1A in the supplemental material. Note that both the TR34 and TR46 promoters contain two SREs and ATREs owing to the 34-bp duplication. The CBC binding site is also duplicated, but the HapX response element (HXRE) is not. To distinguish between these two copies of each site, we refer to them as either the proximal SRE/ATRE (proximal; closest to transcription start [pSRE/pATRE]) or distal SRE/ATRE (distal; furthest from transcription start [dSRE/dATRE]).

Detailed map of cyp51A promoter mutations and analysis of Cyp51A protein levels in response to these alterations. (A) The DNA sequence of the cyp51A promoter region of interest in this study is shown. The wild-type promoter is shown at the top, and the TR34 equivalent is shown at the bottom. Locations of the core binding elements for each transcription are indicated below the DNA sequences. Mutant bases are shown in red lettering for each site. The extent of the 34-bp repeat is shown by the gray highlighting. (B) Whole-cell protein extracts were prepared and analyzed by Western blotting using the anti-Cyp51A antiserum (S. Paul, D. Diekema, and W. S. Moye-Rowley, Eukaryot Cell 12:1619–1628, 2013, https://doi.org/10.1128/EC.00171-13). Strains lacking the pATRE in the wild-type cyp51A promoter context were unable to be grown in the presence of voriconazole and are absent from that analysis. Lanes are numbered at the top of each panel and the numbers near each Cyp51A polypeptide correspond to the quantitation for this experiment. A representative experiment of at least two is shown, and expression levels were all normalized back to the wild-type Cyp51A with no drug treatment. Download FIG S1, PDF file, 0.08 MB.

To evaluate how the ATRE, SRE, and HXRE contribute to expression of both wild-type and TR34 versions of cyp51A, site-directed mutations were constructed in these elements (Fig. S1A) and returned to the natural cyp51A genomic location (Fig. 1A). These strains were tested for their ability to grow in the presence of voriconazole (Fig. 1B), and the level of cyp51A expression was evaluated by reverse transcription-quantitative PCR (RT-qPCR) (Fig. 1C) or using an anti-Cyp51A antibody (Fig. S1B).

FIG 1

Analysis of cyp51A promoter function. (A) Schematic diagram of final structure of wild-type and mutant cyp51A promoter mutants. All mutants analyzed were reintroduced back at the native cyp51A chromosomal location using CRISPR-based recombination with the insertion of the downstream hygromycin cassette as described earlier (17). The relative location of the TR repeat regions is indicated as a box, with the two exons of cyp51A also noted. The hygromycin selection marker (Hyg cassette) is located downstream of the native 3′ end of the cyp51A mRNA. (B) Disk diffusion assay of mutant forms of the cyp51A promoter. For the wild type (wt) and TR34 derivatives, a filter disk containing 0.01 mg of voriconazole was placed in the center of 106 spores of each indicated strain and allowed to grow at 37°C for 72 h. For TR46 derivatives, the voriconazole amount was increased to 0.032 mg. (C) Strains containing the listed versions of either the wild-type (left-hand side) or TR34 (right-hand side) cyp51A gene were grown to mid-log phase with (+) or without (-) voriconazole treatment. Transcriptional behavior of each mutant promoter was assessed by qRT-PCR relative to the tef1 gene. Data are presented for the average of two independent experiments. Numbers above each bar represent the average fold increase for each strain in the presence or absence of voriconazole. Student’s t test was used to compare mRNA changes for the pairs indicated by the lines. Paired conditions were used for the same isolate assayed with and without the drug, while unpaired conditions were used to compare different isolates. Significance is expressed as follows: *, P < 0.05; **, P < 0.01; and ***, P < 0.001.

Loss of either the pSRE (mpSRE) or the pATRE (mpATRE) from wild-type cyp51A caused a slight (mpSRE) or a large (mpATRE) increase in voriconazole susceptibility (Fig. 1B). Removal of the HXRE led to a large decrease in voriconazole susceptibility. These resistance data were fully consistent with the observed expression changes seen by either RT-qPCR measurements (Fig. 1C) or Western blotting (Fig. S1B). Loss of the ATRE from the wild-type cyp51A promoter caused such profound hypersensitivity to voriconazole that we were unable to recover sufficient fungus to assay expression. Together, these data are consistent with both the SRE and ATRE acting as positive regulatory elements and the HXRE acting as a negative element to control cyp51A expression and function.

Insertion of the TR34 promoter into the cyp51A locus led to a decrease in voriconazole susceptibility as seen before (4, 13). Strikingly, loss of either the pSRE or the pATRE from TR34 cyp51A led to a large decrease in voriconazole susceptibility (Fig. 1B). This decrease in voriconazole susceptibility was accompanied by a large increase in the level of Cyp51A expression (Fig. 1C; Fig. S1B). The behavior of each of these proximal element mutations was similar to that caused by loss of the HXRE from TR34 cyp51A. Although these proximal binding sites clearly work as primarily as positive elements in the wild-type promoter context, they appear to be involved in repression in the TR34 promoter, as their loss leads to a large increase in cyp51A expression. Conversely, loss of either of the distal elements (dSRE or dATRE) caused an increase in voriconazole susceptibility, along with a decrease in expression and loss of voriconazole inducibility of cyp51A mRNA (Fig. 1C) consistent with these binding sites acting as positive sites determining TR34 promoter function.

We also produced proximal ATRE and SRE mutant forms of the TR46 cyp51A gene to determine if the unexpected behavior of these elements would extend to this different promoter context. TR46 corresponds to duplication of 46 bases with an identical 5′ endpoint to TR34 and an additional 12 bp at the 3′ end (14). As seen for their counterparts in the TR34 promoter, loss of either the proximal SRE or ATRE caused a decrease in voriconazole susceptibility and an increase in expression compared to the starting TR46 promoter-containing strain.

These data indicate that the increased Cyp51A expression and reduced voriconazole susceptibility caused by the TR34 or TR46 promoter cannot be explained simply by the increased dosage of the duplicated regions present. The proximal and distal regions have distinct behaviors in the TR34 promoter context and likely in the TR46 promoter as well. The distal 34-bp region behaves strictly as a positive regulator of cyp51A transcription, while the proximal element exhibits a negative effect when present in the TR34 promoter. This is quite surprising since loss of the pATRE from the wild-type cyp51A promoter yields a strain that cannot grow in the presence of voriconazole. The same behaviors are seen for the pSRE, although this strain grew, albeit slowly, in the presence of voriconazole.

Given the important role of AtrR in control of cyp51A promoter function, we compared the requirement for this factor in voriconazole resistance and Cyp51A expression in wild-type and isogenic TR34 cyp51A laboratory strains. We also examined the effect of loss of AtrR in two different clinical strains containing either a TR34 promoter-driven cyp51A gene or a TR46 cyp51A locus. Each of the clinical isolates tested is associated with a different mutant form of Cyp51A. The atrR gene was disrupted in all these 4 strains using CRISPR/cas9 and isogenic atrR and atrRΔ derivatives tested for voriconazole susceptibility (Fig. 2A) and expression of Cyp51A by Western blotting (Fig. 2B). Tables 1 and 2 list the strains and plasmids, respectively, used in this study.

FIG 2

AtrR is essential for voriconazole resistance in laboratory and clinical strains. (A) Isogenic atrR and atrRΔ derivatives of the indicated strains were tested for voriconazole resistance by disk diffusion assay. (B) Western blot analysis of the strains listed above was performed using anti-Cyp51A antiserum. Note that the L98H-containing enzymes electrophoresed very close to a nonspecific background signal as we showed earlier (13). The Y121F T289A Cyp51A enzyme has a higher mobility and is clearly resolved below this background polypeptide. (C) Diagram for potential roles of trans- and cis-acting factors at wild-type and TR34 cyp51A promoters. A hypothetical corepressor is pictured that makes multivalent contacts with the key regulators of cyp51A transcription. The proximal ATRE is indicated by a red hatched box. Other binding sites are color-coded with their respective regulators. Azole drugs trigger corepressor dissociation and gene activation. In the case of the TR34 promoter (right-hand diagrams), the distal SRE and ATRE in the upstream 34-bp repeat can bypass corepressor function and activate transcription. The 34-bp (and 46-bp) tandem repeats do not include the HXRE but maintain a CBC binding site. Interaction of CBC with the adjacent HXRE is required for strong binding of these factors (18). Exposure of the TR34 cyp51A gene to azole drugs or loss of the pSRE or pATRE (shown here) triggers strong induction of expression. Induction of expression in the mpATRE TR34 promoter is maximal even in the absence of azole induction. Only TR34 is shown, but we believe that the same mechanisms operate for the TR46 promoter.

TABLE 1

A. fumigatus strains used in this study

Strain	Parent	Genotype	Source or reference
V232-12		TR34 L98H T289A cyp51A	W. Melchers
SPF169	V232-12	atrRΔ::ptrA	This study
V181-51		TR46 Y121F T289A cyp51A	W. Melchers
SPF169	V181-51	atrRΔ::ptrA	This study
AfS35	D141	akuA::loxP	FGSC
SPF92	AfS35	wt cyp51A hph	13
SPF200	AfS35	mSRE cyp51A hph	This study
SPF202	AfS35	mATRE cyp51A hph	This study
SPF204	AfS35	mHXRE cyp51A hph	This study
SPF94	AfS35	TR34 cyp51A hph	13
SPF206	AfS35	mdSRE TR34 cyp51A hph	This study
SPF208	AfS35	mdATRE TR34 cyp51A hph	This study
SPF210	AfS35	mHXRE TR34 cyp51A hph	This study
SPF212	AfS35	mpSRE TR34 cyp51A hph	This study
SPF214	AfS35	mpATRE TR34 cyp51A hph	This study

TABLE 2

Plasmids used in this study

Plasmid	Parent	Genotype	Reference
A1	pUC57	wt cyp51A hph	5
pSP119	A1	mSRE cyp51A hph	This study
pSP120	A1	mATRE cyp51A hph	This study
pSP121	A1	mHXRE cyp51A hph	This study
L5H	pUC57	TR34 cyp51A hph	5
pSP122	A1	mdSRE TR34 cyp51A hph	This study
pSP123	A1	mdATRE TR34 cyp51A hph	This study
pSP124	A1	mHXRE TR34 cyp51A hph	This study
SPF125	A1	mpSRE TR34 cyp51A hph	This study
SPF126	A1	mpATRE TR34 cyp51A hph	This study

The presence of AtrR was essential for the normal high-level voriconazole resistance seen in both clinical isolates, irrespective of the TR34 or TR46 nature of the cyp51A promoter. The overexpression of Cyp51A was also eliminated from these strains when atrR was deleted.

Together, these data illustrate the unexpected complexity of the TR34 promoter region in cyp51A expression. We argue that a simple increase in dosage of a positively acting region of 34 bp cannot explain the unique behavior of the TR34 promoter. The distal 34-bp repeat behaves positively, but the proximal 34-bp repeat has a strong negative effect on TR34 promoter activity. We hypothesize the presence of a multivalent corepressor (Fig. 2C) that must be engaged by SrbA and AtrR, along with CBC and HapX, to normally repress cyp51A transcription. A single transcription factor acting as both a repressor or activator has been extensively documented for mammalian nuclear receptors (15). Loss of the binding sites for SrbA or AtrR strongly activates cyp51A expression in the absence of drug induction but only in the context of a duplication of the cyp51A promoter. Importantly, neither the TR34 or TR46 duplication includes both the CBC and HapX binding sites, suggesting that these must be lost in order to provide the proper context for the upstream repeat to induce cyp51A expression. In the wild-type cyp51A promoter, mutations in either the SRE or the ATRE cannot hyperactivate since these elements are also required for normal expression. AtrR is required for voriconazole resistance and Cyp51A overproduction from TR34 and TR46 promoter-driven cyp51A genes, and as seen earlier with SrbA (16), AtrR is a crucial determinant for azole resistance in clinical isolates of A. fumigatus.

Supplemental materials and methods. Download Text S1, PDF file, 0.04 MB.