Transcriptional Profiling Identifies a Role for BrlA in the Response to Nitrogen Depletion and for StuA in the Regulation of Secondary Metabolite Clusters in Aspergillus fumigatus

Abstract

Conidiation (asexual sporulation) is a key developmental process in filamentous fungi. We examined the gene regulatory roles of the Aspergillus fumigatus developmental transcription factors StuAp and BrlAp during conidiation. Conidiation was completely abrogated in an A. fumigatus ΔbrlA mutant and was severely impaired in a ΔstuA mutant. We determined the full genome conidiation transcriptomes of wild-type and ΔbrlA and ΔstuA mutant A. fumigatus and found that BrlAp and StuAp governed overlapping but distinct transcriptional programs. Six secondary metabolite biosynthetic clusters were found to be regulated by StuAp, while only one cluster exhibited BrlAp-dependent expression. The ΔbrlA mutant, but not the ΔstuA mutant, had impaired downregulation of genes encoding ribosomal proteins under nitrogen-limiting, but not carbon-limiting, conditions. Interestingly, inhibition of the target of rapamycin (TOR) pathway also caused downregulation of ribosomal protein genes in both the wild-type strain and the ΔbrlA mutant. Downregulation of these genes by TOR inhibition was associated with conidiation in the wild-type strain but not in the ΔbrlA mutant. Therefore, BrlAp-mediated repression of ribosomal protein gene expression is not downstream of the TOR pathway. Furthermore, inhibition of ribosomal protein gene expression is not sufficient to induce conidiation in the absence of BrlAp.

Aspergillus species are filamentous fungi with a complex life cycle that is characterized by distinct developmental stages. Development is a highly regulated process that has been studied extensively in the model organism Aspergillus nidulans (3), a minimally pathogenic Aspergillus species. Little is known, however, about the role of these stages of development in the more virulent species Aspergillus fumigatus, which can cause invasive pneumonia and disseminated disease in immunocompromised patients. The asexual life cycle of Aspergillus spp. can be divided into two broad stages. First, conidia (asexual spores) undergo germination to produce filamentous hyphae that grow by elongation. As these hyphae mature, they gain the ability to respond to a variety of stimuli such as nutrient starvation by forming multicellular structures (conidiophores) that produce single cellular conidia and thus begin the cycle again. This stage is termed conidiation, and hyphae that have the capacity to form conidiophores are called developmentally competent.

The process of conidiation is under complex genetic control. In A. nidulans, Timberlake demonstrated that there are in excess of 1,000 distinct mRNAs that are found at increased concentrations during conidiation (33). Despite this staggering number, which was derived from subtractive hybridization experiments, brlA is one of very few genes that have been shown through genetic analysis to be absolutely required for conidiation (1, 7). Further, overexpression of brlA can induce conidiation at times and under conditions in which conidiation does not normally occur (1, 2).

In both A. fumigatus and A. nidulans, the brlA gene encodes a C₂H₂ zinc finger transcription factor that is expressed in mature hyphae upon exposure to conidiation signals. In mutants lacking BrlAp, development arrests at the stalk (an early structure approximately twice the diameter of vegetative hyphae that does not undergo branching) stage of conidiation and the stalks that do form are abnormally long (10), giving the fungus a “bristle” appearance. BrlAp governs the expression of two other genes, wetA and abaA, which together constitute the central regulatory pathway of sporulation (7, 26). Deleting any of these three genes interferes with sporulation and the expression of several developmental mRNAs (7). Although a few BrlAp-dependent genes have been identified in the model organism A. nidulans, the transcriptional program that is governed by BrlAp has not been reported in either A. fumigatus or A. nidulans.

The transcription factor encoded by the stuA gene is also required for the genetic and physical patterning of the conidiophore in A. nidulans (4, 8, 24). Miller et al. have shown that disrupting the A. nidulans stuA gene interferes with the localization of two central regulatory proteins, BrlAp and AbaAp (25), and results in the production of conidiophores that do not have metulae or phialides and in the formation of reduced numbers of conidia that bud directly from vesicles (10). A ΔstuA mutant of A. fumigatus has a similar phenotype (31). Transcriptional profiling has identified several competence-associated, StuAp-dependent genes in A. fumigatus (31), but the role of StuAp in the postcompetence events of conidiation has not been addressed.

Previous studies have found strong associations between conidiation and the production of secondary metabolites (9), which are usually not required for fungal survival in vitro under nonstressful conditions and can have roles as disparate as activating sporulation and producing pigments (9). Additionally, virulence is, in part, believed to be mediated by secondary metabolites whose expression is under developmental control (20, 18). Generally, developmental transcription factors often regulate secondary metabolism. For example, A. fumigatus brlA mutants are defective in the production of some members of the ergot alkaloid class of secondary metabolites, specifically, fumigaclavines A, B, and C (11). The expression of the epipolythiodioxopiperazine (ETP) gliotoxin, which induces host cell apoptosis and is required for normal virulence of A. fumigatus, is dependent on StuAp both in vitro and in vivo (16, 17, 29, 32, 35). However, the role of BrlAp and StuAp in regulating the expression of other secondary metabolites during conidiation has not been examined.

As noted, one of the most powerful stimuli for conidiation in A. fumigatus is nutrient stress. To explore this link between nutrient deprivation and development, we used whole-genome transcriptional profiling and real-time reverse transcription (RT)-PCR to identify BrlAp- and StuAp-dependent genes in the regulatory networks of conidiation, nutrient sensing, and secondary metabolism.

MATERIALS AND METHODS

Strains and growth conditions.

A. fumigatus strain Af293, a clinical isolate, was used as the wild-type strain. The ΔstuA null mutant strain was described previously (31). Construction of the ΔbrlA mutant and the ΔbrlA::brlA complemented strain is described below. All of the strains, with the exception of the ΔbrlA mutant strain, were passaged on YPD agar (1% yeast extract, 2% peptone, 2% glucose, 1.5% agar). The ΔbrlA mutant strain was passaged on YES agar (2% yeast extract, 15% sucrose, 0.05% MgSO₄ · 7H₂O, 1.5% agar).

Disruption and complementation of brlA.

To construct a ΔbrlA null mutant strain, we used the split marker deletion method described by Sheppard et al. (31). A. fumigatus was cotransformed with two DNA constructs, each of which contained a fragment of the hygromycin phosphotransferase (HYG) gene fused to sequences flanking brlA. Each of these constructs contains a region of overlap within the HYG gene which serves as a site for recombination during transformation. When these two constructs were electroporated into A. fumigatus, their homologous integration into the brlA locus reconstructed the HYG gene while deleting the brlA protein coding region.

The two DNA constructs described above were generated by successive rounds of PCR. First, the 5′ (primers F1 and F2) (Table 1) and 3′ (primers F3 and F4) flanking regions of the brlA gene were amplified from genomic DNA, producing brlA flanking sequences fused to a 24-bp M13 sequence at the internal ends of the constructs. In a separate PCR, the HYG gene was amplified from plasmid pAN7-1 by the primer pair M13F-M13R. Finally, fusion PCR with primer pairs F1-HY and YG-F4 was used to join the brlA flanking fragments to the HYG resistance-encoding fragments. A. fumigatus Af293 was then electroporated with 5 μg of each fragment. Nonconidiating, hygromycin-resistant colonies were selected and analyzed for the absence of brlA by PCR.

TABLE 1.

PCR primers used in this study

Primer	Sequence 5′ to 3′
F1	AGCTTGACACGGCCATTTAC
F2	TCCTGTGTGAAATTGTTATCC
	CTGCGCGATTGTCTCTGATTC
F3	GTCGTGACTGGGAAAACCCTGGCG
	CCGCGAGATCAGTATGGAAT
F4	AAGCATGCAGATTGGAAAGC
M13F	CGCCAGGGTTTTCCCAGTCACGAC
M13R	AGCGGATAACAATTTCACACAGGA
HY	GGATGCCTCCGCTCGAAGTA
YG	CGTTGCAAGACCTGCCTGAA
brlA cloning left	AGCTTGACACGGCCATTTAC
brlA cloning right	AAGCATGCAGATTGGAAAGC
p402 left	CTGGCGTAATAGCGAAGAGG
p402 right	GCAGAGCGAGGTATGTAGGC
brlA rev genome
specific	CGCTGTCAGCAGTTGAACAT
brlA rev construct
specific	CCCAGGCGATCAGACATATT
60S acidic ribosomal
protein P2 RT left	TCTCTTCCGTTGGCATTGAT
60S acidic ribosomal
protein P2 RT right	GAAGCGAGCTTGGTGGAAC
40s ribosomal protein
RT left	CGGCTATGTCAAGACCCAGT
40s ribosomal protein
RT right	ACTCGCGGAGGTAGTCAAGA
Ribosomal L22e protein
family RT left	TCACATCCCCTTCTCTGGTC
Ribosomal L22e protein
family RT right	GCTCGTAGACACCCTTGGAG
TEF1 RT sense	CCATGTGTGTCGAGTCCTTC
TEF1 RT antisense	GAACGTACAGCAACAGTCTGG

To complement the ΔbrlA mutant with the single copy of wild-type brlA, a DNA fragment encompassing the intact brlA open reading frame, as well as 2.3 kb of the sequence upstream of the brlA gene and 1 kb of the downstream sequence, was amplified from genomic DNA of Af293 with the primer pair brlA cloning left and right (Table 1) by high-fidelity PCR. This insert was then cloned into phleomycin resistance plasmid p402 at the unique NotI site. The resulting plasmid was then linearized at its SacI site. Protoplasts of the ΔbrlA mutant were transformed with this construct by a minor modification of the previously described method (6). The concentrations of driselase, β-d-glucanase, and lyticase in the 2× protoplasting solution were increased to 1.25%, 2.4%, and 187.5 U/ml, respectively. In addition, fungal protoplasting was performed for 4 h. Mutants that were resistant to both hygromycin and phleomycin (consistent with the integration of both the disruption and complementation cassettes) were selected. Integration of the intact brlA allele at its native locus was confirmed by PCR with the primers named brlA rev genome specific and brlA rev construct specific shown in Table 1.

Slide culture and microscopy.

To examine the morphological effects of brlA deletion, slide cultures were prepared; fungi were grown on YPD agar sandwiched between glass coverslips at 37°C. After 3 days, the hyphae were stained with lactophenol cotton blue and visualized by light microscopy at a magnification of ×40.

Microarray RNA time course.

To identify genes whose expression was dependent on BrlAp or StuAp, we performed whole-genome transcriptional analysis studies during conidiation. RNA was extracted from wild-type strain Af293, the ΔstuA null mutant, the ΔbrlA null mutant, and the ΔbrlA::brlA complemented strain as detailed below. Since we have previously demonstrated genome-wide specific complementation of the ΔstuA mutant strain (31), the stuA-complemented strain was not included in this study.

When grown submerged in rich media such as YPD medium, A. fumigatus can be maintained in a state of developmental competence without progression to conidiation (31). Because the ΔbrlA mutant was aconidial, it was always present as competent hyphae. Thus, to match all of the strains to a similar point in the developmental cycle before transcriptional analysis, we synchronized them in a state of developmental competence in liquid YPD medium and performed the conidiation time course study beginning with these synchronized cultures. Briefly, hyphal fragments from two plates of the ΔbrlA mutant were collected with a cotton swab and inoculated into 160 ml of YPD medium. In parallel, 1.6 × 10⁵ conidia of the Af293, brlA-complemented, and ΔstuA strains were inoculated into separate flasks, each containing 160 ml of YPD medium. All of the strains were grown for 30 h at 37°C in a shaking incubator. Next, the flasks were incubated for a further 24 h at room temperature without agitation after the addition of another 60 ml of YPD medium to ensure that the cultures remained submerged and did not conidiate. Then, in order to stimulate conidiation, hyphae were transferred to RPMI 1640 medium (Sigma-Aldrich product number R6504) buffered at pH 7.0 with 34.5 g/liter 3-(N-morpholino)propanesulfonic acid (MOPS; Sigma-Aldrich) and incubated at 37°C in a shaking incubator. The same amount of hyphae was added to each flask, as assessed by visual inspection and comparison of filter-collected hyphal mats. After 1, 8, 16, and 24 h of growth in MOPS-buffered RPMI 1640 medium, an aliquot of each strain was removed and RNA was extracted with the Qiagen RNeasy Plant Mini Kit by following the manufacturer's protocol.

Microarray analysis.

A DNA amplicon microarray for A. fumigatus Af293 (28) was used to determine relative gene expression levels. Each gene was present in triplicate on the array, and all hybridizations were repeated in dye swap experiments. The data for each gene were averaged from the triplicate genes on each array and the duplicate dye swap experiment (a total of six readings for each gene). Genes that exhibited significantly different expression levels compared to the 1-h reference time point for each strain (Z score, >1.96), were analyzed by Euclidean distance and hierarchical clustering by the average linkage clustering method in the multiple-experiment viewer of The Institute for Genomic Research as described previously (31). Two independent biological replicates performed on separate days were analyzed, and two supplementary independent replicates were used for real-time RT-PCR verification experiments as detailed below.

Real-time RT-PCR.

We examined the expression levels of four genes (brlA and three ribosomal genes) by real-time RT-PCR. RNA was isolated at the time points described above. RNA was DNase treated on the column and reverse transcribed with Moloney murine leukemia virus reverse transcriptase (Invitrogen) according to the manufacturer's instructions. After RT, a real-time PCR was performed with an ABI 7300 thermocycler and the primers designated for RT in Table 1. PCR products were detected by the Sybr green method, and gene expression was normalized to the gene TEF1 to obtain threshold cycle (ΔC_T) values (C_{Ttarget gene} − C_TTEF1). Graphs show relative expression levels obtained with the formula 2^−ΔΔCT, where ΔΔC_T = ΔC_{Ttarget gene} − ΔC_Tcalibrator and the calibrator is a strain and/or time point where the gene expression level has been arbitrarily set to a value of 1 as described in each experiment. As negative controls for each gene, samples were run which had not been reverse transcribed to ensure the absence of genomic DNA contamination.

Target of rapamycin (TOR) inhibition RNA time course.

Hyphae were synchronized at the stage of developmental competence as previously described. Equal amounts of hyphae were inoculated into 50-ml conical tubes containing either 20 ml of YPD medium or 20 ml of YPD medium plus 100 ng/ml rapamycin (Bioshop, solubilized in dimethyl sulfoxide) and incubated with shaking at 37°C. RNA was extracted at 30 minutes and at 4 hours after inoculation with the Machery Nagel Nucleospin RNA plant extraction kit by following the manufacturer's protocol and gene expression was analyzed by real-time RT-PCR.

Nutrient starvation RNA time course.

Equal amounts of synchronized hyphae were inoculated into 50-ml conical tubes containing 20 ml of Aspergillus minimal medium, Aspergillus minimal medium lacking glucose (carbon deficient), or Aspergillus minimal medium without NaNO₃ (nitrogen deficient), RNA was extracted at 30 minutes and at 4 hours after inoculation, and gene expression was analyzed as described above.

RESULTS

Morphological abnormalities in the ΔbrlA null mutant.

The morphological effects of brlA deletion were examined macroscopically and by slide culture (Fig. 1). The ΔbrlA mutant strain was not pigmented and produced no structures beyond the stalk stage of differentiation. Complementation of the ΔbrlA mutant strain with a wild-type copy of the brlA gene fully restored the ability to produce conidiophores.

FIG. 1.

The ΔbrlA mutant grows as constitutive hyphae. (Top left) Hyphae were plated on YES agar and grown for 3 days at 37°C. (Top right) Wild-type A. fumigatus hyphae and conidiophores. (Bottom left) The ΔbrlA mutant produces no differentiated structures. (Bottom right) Complementation (com.) of the ΔbrlA mutant with a wild-type copy of the gene restores normal conidiation.

Microarray analysis reliably identifies transcriptionally regulated genes.

To confirm that our experimental approach was consistent with previously studied developmental time courses, we examined the expression of genes that have been previously identified as BrlAp dependent in A. nidulans (7, 19, 23, 34, 36). Additionally, we examined the expression of genes known to be involved in conidiation and pigment biosynthesis since the ΔbrlA mutant strain produces no conidia and is not pigmented. As shown in Fig. 2, the majority of these candidate genes were upregulated in strain Af293 and the brlA-complemented strain but not in the conidiation-deficient ΔbrlA or ΔstuA null mutant strain. Consistent with these findings, conidiation was visually evident in the wild-type and brlA-complemented strains after 16 h of growth in RPMI medium. Collectively, these data suggest that this experimental protocol appropriately recapitulates the developmental program of conidiation.

FIG. 2.

Microarray confirmation of known BrlAp-dependent genes. Selected genes from the microarray RNA time course study were examined to ensure that their regulation was concordant (BrlA dependent) with gene regulation which is known through genetic analyses. Green, upregulation; red, downregulation; gray, no data; w.t., wild type; ΔbrlA, brlA deletion mutant; com., brlA-complemented strain; ΔstuA, stuA deletion mutant. In this and subsequent microarray figures, the color of each spot represents the log₂-fold change in gene expression over the 1-h time point for the same strain.

BrlAp and StuAp direct distinct transcriptional programs during conidiation.

Next, we examined the relative contributions of BrlAp and StuAp to the regulation of the A. fumigatus transcriptome during conidiation. At 8 and/or 16 h of growth, 568 genes were significantly altered in expression (Z score, >1.96) in wild-type strain Af293 (346 upregulated and 222 downregulated), compared with 1 h of growth. In the brlA-complemented strain, 575 genes showed significant differences in expression level (411 upregulated and 164 downregulated). Of these genes, 374 were significantly differentially expressed in both Af293 and the brlA-complemented strain (287 up and 87 down). For a comprehensive list of these genes, see Table S1 in the supplemental material; a subset of these genes are presented in detail in Tables 2 and 3, organized by putative gene product and function. For the complete data set, see Table S2 in the supplemental material.

TABLE 2.

Subset of genes upregulated in Af293 and the brlA-complemented strain

Common name	Locus	Siga in ΔbrlA strain	Sig in ΔstuA strain
Regulation of development
Meiosis induction protein kinase (Ime2), putative	Afu2g13140	Yes	Yes
Polarized growth protein (Boi2), putative	Afu2g03360	Yes	Yes
Sexual development transcription factor NsdD	Afu3g13870	Yes	Yes
AbaA protein	Afu1g04830	Nob	Nob
Regulatory protein WetA	Afu4g13230	No	No
Cell wall protein (PhiA), putative	Afu3g03060	Nob	No
Secondary metabolism
Nonribosomal peptide synthase, putative	Afu6g09610	Yes	Yes
Cytochrome P450, putative	Afu6g14360	Yes	No
Mycelial catalase Cat1	Afu3g02270	Yes	No
P450 monooxygenase, putative	Afu6g13940	Yes	No
Cytochrome P450 monooxygenase, putative	Afu4g14790	No	No
Dimethylallyl tryptophan synthase, putative	Afu2g18040	No	No
Polyketide synthase, putative	Afu8g00370	No	No
Phenol 2 monooxygenase, putative	Afu6g03490	No	No
Antigens/cell surface proteins
Major allergen Asp f1	Afu5g02330	Yes	Yes
Major allergen Asp f2	Afu4g09580	Yes	Yes
Allergen Asp f13	Afu2g12630	No	Yes
IgEc-binding protein	Afu6g00430	No	Yes
Transport
ABC multidrug transporter SitT, putative	Afu3g03430	Yes	Yes
High-affinity iron permease CaFTR2	Afu5g03800	Yes	Yes
MFSd toxin efflux pump, putative	Afu6g02220	Yes	Yes
MFS siderophore transporter, putative	Afu3g03440	Yes	Yes
Ctr copper transporter family	Afu6g02810	Yes	No
Nonclassical export protein (Nce2), putative	Afu2g01590	Yes	No
Small oligopeptide transporter, OPT family	Afu3g12200	Yes	No
ABC multidrug transporter, putative	Afu6g03470	No	Yes
Metabolism
6-Phosphofructo-2-kinase 1	Afu1g07220	Yes	Yes
Short-chain dehydrogenase, putative	Afu4g08710	Yes	Yes
Trehalose synthase (Ccg-9), putative	Afu3g12100	Yes	No
UDP-glucose dehydrogenase Uxs2p	Afu8g00920	Yes	No
Cholesterol delta-isomerase, putative	Afu3g00810	Yes	Yes
Glutamate carboxypeptidase	Afu2g04370	No	No

^aSig, whether or not a gene was significantly upregulated in the indicated mutant strain.

^bThis gene was actually downregulated in this mutant.

^cIgE, immunoglobulin E.

^dMFS, major facilitator superfamily.

TABLE 3.

Subset of genes downregulated in Af293 and the brlA-complemented strain

Common name	Locus	Siga in ΔbrlA	Sig in ΔstuA
Ribosomal genes
60S ribosomal protein L12	Afu1g03390	No	Yes
60S ribosomal protein L27	Afu1g06340	No	Yes
60S ribosomal protein L29, putative	Afu1g03110	No	Yes
60S ribosomal protein L35, putative	Afu1g10510	No	Yes
Ribosomal L38e protein family	Afu4g07830	No	Yes
Ribosomal protein P0	Afu1g05080	No	Yes
Ribosomal protein S10	Afu2g02150	No	Yes
Secondary metabolism
Cytochrome P450, putative	Afu6g02210	Yes	Yes
Nonribosomal peptide synthase, putative	Afu3g03350	Yes	Nob
Antigens/cell surface proteins
Antigenic dipeptidyl-peptidase Dpp4	Afu4g09320	Yes	No
Cell surface protein Mas1, putative	Afu8g00610	Nob	Yes
GPIc-anchored protein, putative	Afu8g04370	Yes	No
Transport
Arginine transporter, putative	Afu5g04260	Yes	Yes
Amino acid permease (Dip5), putative	Afu2g08800	Yes	Yes
MFSd sugar transporter Stl1, putative	Afu8g05710	Yes	Yes
Uracil permease	Afu1g13210	Yes	Yes
Major facilitator superfamily	Afu5g01350	Yes	Yes
Proline permease	Afu8g02200	Yes	Nob
Ammonium transporter MeaA	Afu2g05880	Yes	No
Metabolism
Oxidoreductase, short-chain dehydrogenase/reductase family	Afu5g11240	Yes	Yes
Aldehyde dehydrogenase, putative	Afu6g11430	Yes	Yes
Aldehyde dehydrogenase, putative	Afu7g01000	Yes	Yes
Alcohol dehydrogenase, putative	Afu7g01010	Yes	Nob
Aldehyde reductase (AKR1), putative	Afu6g10260	No	No
Glutamate carboxypeptidase, putative	Afu3g05450	Yes	No
Oxidoreductase, short-chain dehydrogenase/reductase family	Afu6g09140	Yes	No
Diacylglycerol O-acyltransferase (DgaT), putative	Afu2g08380	No	Nob
Adenine phosphoribosyltransferase 1	Afu7g02310	Yes	Nob
Methyltransferase	Afu8g01930	Yes	No

^aSig, whether or not a gene was significantly downregulated in the indicated mutant strain.

^bThis gene was actually upregulated in this mutant.

^cGPI, glycosylphosphatidylinositol.

^dMFS, major facilitator superfamily.

We next determined which of these conidiation-associated genes were differentially regulated in the ΔbrlA and/or ΔstuA null mutant strains. Thirty of the conidiation-associated genes were uniquely dysregulated in the ΔbrlA mutant, while 108 were uniquely dysregulated in the ΔstuA mutant and 105 genes displayed patterns of altered regulation in both mutants. A Venn diagram illustrating the relationship between BrlAp- and StuAp-dependent conidiation-associated genes is shown in Fig. 3, and a subset are further detailed in Tables 2 and 3. As is evident from these data, all combinations of differential regulation were observed (i.e., dependent on neither BrlAp nor StuAp, dependent on both, and only dependent on one). Two striking examples of genes exhibiting differential regulation were the expression of genes involved in secondary metabolism, which show a strong bias toward StuAp dependence, and downregulation of the expression of ribosomal genes, which were BrlAp dependent. Collectively, these data indicate that despite their common role as developmental regulators, BrlAp and StuAp direct distinct transcriptional programs during conidiation. In the results and discussion that follow, we describe in more detail the examples of gene regulation of toxin production and ribosomal protein expression during conidiation.

FIG. 3.

BrlAp and StuAp direct distinct but overlapping transcriptional programs during conidiation. Two hundred forty-three genes were dysregulated in the mutant strains during conidiation. The distribution of these genes is represented by a Venn diagram. Each number in parentheses is the number of genes in a particular category. Black, genes that are uniquely dysregulated in the ΔbrlA mutant; white, genes that are uniquely dysregulated in the stuA mutant; gray, genes that are dysregulated in both the ΔstuA and ΔbrlA mutant strains.

Dependence of secondary metabolism on developmental transcription factors.

In fungi, development metabolism and secondary metabolism are often tightly associated. Therefore, we examined the role of BrlAp and StuAp in the regulation of the expression of secondary metabolite genes. Significant differences in the expression of secondary metabolite biosynthetic clusters were found between strains during conidiation (Fig. 4). Six of the 22 putative biosynthetic clusters were identified whose expression was StuAp dependent. Two of the clusters are involved in the synthesis of alkaloids (fumigaclavine and fumitremorgen). Another two clusters are of the ETP class (gliotoxin and an unknown ETP-like toxin). One cluster is predicted to produce pseurotin A, and the product of the last cluster is unknown. In contrast, only the fumigaclavine-producing cluster exhibited BrlAp-dependent regulation.

FIG. 4.

Secondary metabolism is dependent upon developmental transcription factors. Shown are heat maps of biosynthetic clusters encoding enzymes required for the synthesis of fumitremorgen (Ft), pseurotin A (Pt), gliotoxin, a putative ETP molecule, and fumigaclavine, an ergot alkaloid. Tn, retrotransposon; ?, unknown molecule; w.t., wild-type strain; com., complemented strain.

Ribosomal gene expression during conidiation is BrlAp, but not StuAp, dependent.

An unexpected finding of this study was that the regulation of ribosomal protein expression during conidiation was BrlAp dependent. In the wild-type, brlA-complemented, and ΔstuA strains, the shift to RPMI 1640 medium and conidiation was associated with a coordinated downregulation of the expression of more than 70 genes encoding ribosomal proteins. This response began at 8 h in these strains and was complete by 16 h of growth. In contrast, in the ΔbrlA strain, the expression of these genes remained at basal levels, and in some cases the genes were even upregulated, until 24 h after a shift to conidiation-inducing conditions (Fig. 5). We verified these results by real-time RT-PCR analysis of the expression of three ribosomal proteins (Fig. 6). At 16 h of growth, mRNA expression levels of all three of the ribosomal proteins tested were significantly lower than the expression levels in the ΔbrlA mutant strain. Collectively, these results suggest that BrlAp may play a role in mediating the temporal expression of ribosomal protein synthesis.

FIG. 5.

Ribosomal gene expression during conidiation is BrlAp but not StuAp dependent. Labels and data analysis are as described for Fig. 2. w.t., wild-type strain; com., complemented strain.

FIG. 6.

Real-time PCR confirms the microarray-obtained ribosomal gene expression data. To confirm the microarray results, a real-time PCR was performed with two independent sets of RNA isolated from each strain. Gene expression was normalized to A. fumigatus gene TEF1, and the expression of each gene after 16 h of growth in RPMI medium was also normalized relative to the respective strain at 1 h and is shown on the y axis. Error bars represent the standard error of duplicate samples. An asterisk indicates significantly reduced gene expression compared to the strain Af293 sample at the same time (P < 0.05 by factor analysis of variance).

BrlAp is required for ribosomal protein gene suppression during nitrogen, but not carbon, starvation.

Since both conidiation and downregulation of ribosomal protein biogenesis are elements of the starvation response in fungi and RPMI 1640 medium is a relatively nutrient-poor medium (only 2 g/liter of d-glucose for a 1-g/liter mixture of 20 amino acids), we hypothesized that BrlAp may play a role in the response to starvation. To further explore the role of BrlAp in nutrient availability responses, the expression of three ribosomal genes was monitored under conditions of either carbon or nitrogen starvation. Transfer of hyphae to medium deficient in glucose resulted in a significant decrease in ribosomal protein mRNA levels for all three of the genes tested (Fig. 7). This decrease was seen in both the wild-type and brlA mutant strains, suggesting that brlA is not required to mediate the suppression of ribosomal protein gene expression in response to carbon starvation. In contrast, under conditions of nitrogen starvation, differences in ribosomal protein expression were seen between the wild-type and brlA mutant strains (Fig. 7). In response to 4 h of nitrogen starvation, significantly lower levels of ribosomal protein expression were observed in wild-type strain Af293 than in the ΔbrlA mutant strain. These results suggest that brlA may mediate the suppression of ribosome biogenesis during nitrogen, but not glucose, starvation.

FIG. 7.

BrlAp is required for downregulation of ribosomal protein gene expression in response to nitrogen starvation. A real-time PCR was used to test the response of A. fumigatus ribosomal protein gene expression during carbon and nitrogen starvation. The expression level of each gene in each strain after 4 h of nutrient deprivation was normalized both to A. fumigatus TEF1 and to the level of gene expression in nutrient-replete medium for the same strain. Error bars represent the standard error of four readings (duplicate samples from two independent experiments). An asterisk indicates significantly increased gene expression compared to the strain Af293 sample (P < 0.05 by factor analysis of variance). MM, minimal medium.

BrlAp-mediated ribosomal protein gene suppression is TOR independent.

Starvation responses in fungi such as Saccharomyces cerevisiae can be mediated through TOR pathway signaling. To determine if BrlAp-induced ribosomal protein suppression was mediated through TOR signaling, we evaluated the effects of rapamycin inhibition of TOR on both ribosomal biogenesis and conidiation. Exposure of wild-type hyphae of strain Af293 growing in rich medium to rapamycin resulted in the induction of conidiation, despite growth in submerged culture and under nutrient-replete conditions. This conidiation was less marked than the conidiation seen with RPMI medium induction or nitrogen starvation. In parallel, an increase in brlA expression and a reduction in ribosomal protein gene expression were observed (Fig. 8). Consistent with the observation that only a moderate level of conidiation was observed, the changes in brlA and ribosomal protein gene expression were less marked than the changes seen in response to growth in RPMI medium (Fig. 6). Exposure of the ΔbrlA mutant strain to rapamycin produced a similar reduction in ribosomal protein gene expression, although conidiation was not induced. Collectively, these results suggest that either (i) the observed BrlAp-dependent suppression of ribosomal protein gene expression is TOR independent or (ii) BrlAp functions upstream of TOR. Furthermore, suppression of ribosomal gene expression is not sufficient to induce conidiation in the absence of BrlAp.

FIG. 8.

Rapamycin-mediated ribosomal protein gene suppression does not require BrlAp. A real-time PCR was used to examine the effects of TOR inhibition (growth for 4 h in YPD medium supplemented with 100 ng/ml rapamycin [Rap.]) on ribosomal protein expression. Gene expression was normalized to A. fumigatus TEF1 and relative to expression under the YPD medium condition, as shown. Error bars represent the standard error of four readings (duplicate samples from two independent experiments). An asterisk indicates significantly reduced gene expression in YPD medium plus rapamycin at 4 h of growth compared to that of the same strain in YPD medium at the same time point (P < 0.05 by factor analysis of variance).

DISCUSSION

Conidiation is associated with dramatic morphological changes and is critical for the dissemination of fungal propagules within the environment. Although it is recognized that development in Aspergillus species is remarkably complex, many of the genes and pathways involved in the regulation of development remain unidentified. We used whole-genome microarray analysis to investigate the abnormalities in gene expression in two A. fumigatus mutants (ΔbrlA and ΔstuA) in which the process of conidiation is markedly impaired.

Since conidiation is abnormal in both the ΔbrlA and ΔstuA mutants, one might predict that the transcriptional programs governed by these two transcription factors are similar. However, brlA encodes a core essential regulator of development while stuA encodes a so-called modifier of development. Thus, conidiation still occurs in the ΔstuA mutant, albeit abnormally. In keeping with the related yet unique phenotypes of these two mutants, we observed that BrlAp and StuAp direct distinct yet overlapping transcriptional programs during conidiation. In total, we identified 243 genes that were differentially regulated between the wild-type strain and one or both of the conidiation-deficient mutants. Only 30 of these were exclusively dependent on BrlAp, in contrast to the 108 that were exclusively dependent on StuAp, while 105 were dependent upon both transcription factors for their proper expression. Therefore, although there is significant overlap, it is clear that BrlAp and StuAp also have independent functions. One limitation of these studies, however, is that they cannot provide an exhaustive list of the BrlAp- and StuAp-dependent genes involved in conidiation since the differential regulation of some genes may be of biological significance but fail to meet the criteria for statistical significance. Nevertheless, these findings provide an important starting point for our understanding of this complicated developmental process.

Past studies have found a correlation between development and secondary metabolite production (9). We found that six different secondary metabolite gene clusters showed a strong dependence on StuAp for the expression of their genes. Of these six, three are contained within an unusual “supercluster” (Afu8g00100 to Afu8g00720) on chromosome 8 (30). This region is believed to encompass three distinct clusters responsible for the synthesis of fumitremorgens (21), pseurotin A (22), and a third, unknown, metabolite (30). Our data support the presence of three distinct clusters within the supercluster, since there are three distinct patterns of regulation in this region. The fumitremorgen cluster is StuAp dependent and suppressed by BrlAp, while the other two clusters are BrlAp independent and StuAp dependent but display a clear difference in the intensity of their upregulation. Additionally, the presence of a retrotransposon between the fumitremorgen and pseurotin A clusters raises the possibility that these clusters were joined by retrotransposon-mediated fusion. The boundaries of this supercluster, as determined by our expression data, are in agreement with a previous report which found the supercluster to also be dependent on the transcription factor LaeA (30). Interestingly, only one of the clusters (encoding the genes required for the synthesis of the ergot alkaloid fumigaclavine) exhibited BrlAp dependence for the expression of its genes. These data are consistent with the observation that the ΔbrlA mutant strain is deficient in the production of ergot alkaloids (11). Since conidiation is abrogated in the ΔbrlA strain yet secondary metabolism is relatively unaffected, these data highlight that conidiation is not an absolute requirement for the elaboration of many of the products of secondary metabolism. This finding is of clinical relevance given that hyphae are the only form of the organism present during invasive A. fumigatus infection and conidiation does not occur (except in pulmonary mycetomas).

One unexpected finding of this study was that elements of the nutrient stress response appeared to be BrlAp dependent. During conidiation, we observed that the expression of almost 80 ribosomal protein genes was coordinately suppressed in all of the strains except the ΔbrlA mutant. This decrease in ribosome biogenesis suggests that during conidiation there is an overall metabolic switch from active vegetative growth to a lower energetic state. These findings are consistent with the existing view that starvation provides an important signal for reducing vegetative growth and the induction of sporulation to enhance escape from unfavorable environmental conditions such as growth in nutrient-poor RPMI medium (5, 12, 13, 15, 27). As would be expected, the ΔbrlA mutant strain was defective in sporulation in RPMI medium, consistent with its known role in governing the core program of conidiation. Surprisingly, the ΔbrlA mutant was also impaired in the inhibition of ribosomal protein gene expression in response to this starvation stimulus. These results suggest a link between conidiation and ribosomal protein biogenesis that is downstream of BrlAp. Thus, BrlAp is not simply a regulator of the production of reproductive structures but also can influence the transition from vegetative growth to metabolic inactivity during conidiation. While we cannot exclude a direct role for BrlAp in the control of ribosomal protein RNA production, it seems more likely that this regulation is indirect, perhaps through the regulation of other secondary signaling molecules.

We found that ribosomal protein mRNA expression in the ΔbrlA mutant was insensitive to nitrogen starvation whereas it responded normally to carbon source deprivation. The TOR signaling pathway is involved in the regulation of metabolic responses to nutrient availability such as translation of mRNAs, ribosome biogenesis, and the regulation of various permeases for sugars and amino acids (5).

Indeed, treatment of wild-type A. fumigatus with the TOR inhibitor rapamycin induced conidiation and brlA gene expression and reduced ribosomal protein mRNA expression, even in a rich medium. However, this suppression of ribosomal protein expression was less dramatic than that observed with RPMI medium starvation. Further, deletion of brlA had no effect on ribosomal protein gene expression upon rapamycin exposure, suggesting that while TOR does indeed play a role in regulating the switch between vegetative and reproductive growth, BrlAp does not function downstream of TOR in regulating ribosomal protein gene expression. In contrast, brlA deletion abrogated rapamycin-mediated conidiation, suggesting that TOR-mediated conidiation signals lie upstream of BrlAp. Collectively, these results highlight the complex nature of the regulatory pathways governing conidiation and developmental growth arrest.

Fungal ribosome biogenesis in response to nutrient starvation is also controlled by the protein kinase A (PKA)-cyclic AMP (cAMP) pathway in fungi (5). Thus, it is possible that BrlAp functions downstream of the PKA-cAMP pathway in the inhibition of ribosome biogenesis. In support of this hypothesis, Fortwendel et al. found that dominant negative A. fumigatus rasB mutants displayed premature expression of brlA mRNA (14). This increase in brlA mRNA levels was associated with the formation of conidiophores in submerged culture, while no conidiation was seen in wild-type strain Af293. However, the effects of dominant negative rasB mutation on ribosomal protein biogenesis were not studied. Examining the effects of brlA deletion on ribosomal protein inhibition in the background of a dominant negative rasB mutation might provide insights into the role of BrlAp in the PKA-cAMP pathway. These and other studies are required to clearly delineate the relationship of BrlAp and ribosomal biogenesis at a molecular level.

In summary, this study provides the first report of the transcriptome during conidiation in an Aspergillus species and shows the relative contributions of two crucial transcription factors to this process. Our data provide evidence that StuAp has a key role in regulating secondary metabolism and that BrlAp has a key role in regulating nitrogen stress responses. Our transcriptional profiling data provide a foundation for future genetic and biochemical analyses that are required to elucidate the mechanisms of and links between the control of conidiation, secondary metabolism, and nutrient sensing in A. fumigatus.