Laccases Involved in 1,8-Dihydroxynaphthalene Melanin Biosynthesis in Aspergillus fumigatus Are Regulated by Developmental Factors and Copper Homeostasis

Abstract

Aspergillus fumigatus produces heavily melanized infectious conidia. The conidial melanin is associated with fungal virulence and resistance to various environmental stresses. This 1,8-dihydroxynaphthalene (DHN) melanin is synthesized by enzymes encoded in a gene cluster in A. fumigatus, including two laccases, Abr1 and Abr2. Although this gene cluster is not conserved in all aspergilli, laccases are critical for melanization in all species examined. Here we show that the expression of A. fumigatus laccases Abr1/2 is upregulated upon hyphal competency and drastically increased during conidiation. The Abr1 protein is localized at the surface of stalks and conidiophores, but not in young hyphae, consistent with the gene expression pattern and its predicted role. The induction of Abr1/2 upon hyphal competency is controlled by BrlA, the master regulator of conidiophore development, and is responsive to the copper level in the medium. We identified a developmentally regulated putative copper transporter, CtpA, and found that CtpA is critical for conidial melanization under copper-limiting conditions. Accordingly, disruption of CtpA enhanced the induction of abr1 and abr2, a response similar to that induced by copper starvation. Furthermore, nonpigmented ctpAΔ conidia elicited much stronger immune responses from the infected invertebrate host Galleria mellonella than the pigmented ctpAΔ or wild-type conidia. Such enhancement in eliciting Galleria immune responses was independent of the ctpAΔ conidial viability, as previously observed for the DHN melanin mutants. Taken together, our findings indicate that both copper homeostasis and developmental regulators control melanin biosynthesis, which affects conidial surface properties that shape the interaction between this pathogen and its host.

INTRODUCTION

Aspergillus fumigatus is an opportunistic pathogen that causes life-threatening invasive disease in immunocompromised hosts (1). Its asexual spores, named conidia, are the initial source of inocula for infection and are the vehicles for dispersal and survival. Under favorable conditions, conidia break dormancy and grow as elongated, highly polarized hyaline hyphae. Upon achieving competency and exposure to air, hyphae produce aerial conidiophore stalks and develop elaborate multicellular conidiophores that generate chains of melanized uninucleate conidia (2, 3) (Fig. 1A and B). BrlA, a C₂H₂ zinc finger transcription factor, is the master regulator that controls the initiation of conidiophore development in Aspergillus (4, 5). AbaA and WetA, two other important regulators that function downstream of BrlA, are required for proper progression of conidiation (6). The abaA gene is activated during the middle stage of conidiophore development, while the wetA gene is required for the activation of late conidiation-specific genes.

Fig 1.

Different developmental stages of A. fumigatus. (A) Asexual reproductive life cycle of A. fumigatus. (B) Melanization increases with conidial maturation. Black arrows point to mature conidiophores with a darker color, and white arrows point to young conidiophores with a lighter color. (C) A. fumigatus of different developmental stages collected at the following time points: 12 h (vegetative hyphae) (a), 24 h (competent hyphae) (b), 36 h (early conidiophores) (c), and 48 h (conidiophores with mature conidia) (d). (e, f, and g) Thirty-six-hour sample at higher magnification. (h and i) Forty-eight-hour sample at higher magnification. Scale bars, 10 μm (b and d) and 5 μm (i); each scale bar shown applies to all panels within the same row of images.

The 1,8-dihydroxynaphthalene (DHN) melanin coating A. fumigatus conidia gives them their characteristic bluish green color (Fig. 1B). Melanin is an amorphous polymer that is produced by a variety of microbes and higher eukaryotes. Melanin not only helps fungi to fend off various environmental insults (e.g., UV irradiation) but also helps the survival in hosts of different animal and plant pathogens such as A. fumigatus, Botrytis cinerea, Cryptococcus neoformans, Exophilia dermatitidis, and Magnaporthe grisea (7–13). Melanin is required for full virulence in Cryptococcus, and melanized Cryptococcus cells are found in infected host tissues (8, 14–17). Many phytopathogenic fungi, such as Colletotrichum lagenarium, Colletotrichum lindemuthianum, and Magnaporthe grisea, require melanized appressoria to penetrate host cells (18–20).

The DHN melanin and the l-3,4-dihydroxyphenylalanine (l-dopa) melanin are the two major types of melanin synthesized by fungi. The DHN melanin is synthesized de novo using endogenous substrates through the DHN intermediate. The final product is produced by a series of enzymatic reactions involving a polyketide synthase and other enzymes such as scytalone dehydratase, hydroxynaphthalene reductase, and laccases (21, 22). The genes encoding these enzymes are organized in a cluster in some fungi, such as A. fumigatus, Aspergillus niger, Aspergillus clavatus, and Cochliobolus heterostrophus (22, 23). However, such clustered organization of melanin biosynthetic genes is not conserved even among some close relatives such as Aspergillus flavus, Aspergillus nidulans, and Aspergillus terreus (21, 24). In contrast, the l-dopa melanin is synthesized by laccases, and its production requires exogenously supplied phenolic precursors such as l-dopa (15, 16, 21). l-Dopa melanin is the well characterized in C. neoformans, and the deletion of the laccase gene attenuates cryptococcal virulence in animal models (8, 17).

Regardless of the organisms or the type of melanin (DHN or l-dopa), laccases are the required and conserved factors for melanin biosynthesis. Although laccases have been among the earliest and most intensively studied enzymes, factors that regulate their activities are not defined in the vast majority of microbes. What is clear is that melanization is associated with development in some filamentous fungi. For instance, in the model filamentous fungus A. nidulans, the laccase level drastically increases during conidiation (25). Consistently, the expression of the wA gene encoding a polyketide synthase and that of the yA gene encoding a p-diphenol oxidase in A. nidulans are both upregulated during conidiation (22, 26). Similarly, in A. fumigatus, the DHN gene cluster, including the laccase-encoding abr1 and the abr2 genes, is expressed during conidiation but not in vegetative hyphae (22, 27–29). However, the role of the central regulators of conidiation (BrlA, AbaA, and WetA) in melanization is still unclear.

The importance of copper in melanization has been highlighted in a few studies; this importance is likely due to the requirement of copper as a cofactor for laccases (30, 31). In A. nidulans, a mutation in the ygA gene that encodes a transporter produces yellow conidia at high pH and green conidia at low pH (32, 33). The defect in melanization of the ygA mutant is likely caused by a deficiency in copper uptake, as the defect can be remediated by the addition of copper (32). Similarly, in Colletotrichum lindemuthianum, the disruption of the copper transporter clap1 renders the strain nonpathogenic and causes poor pigmentation in the spores. Such defects, again, can be remediated by the addition of copper (34). In humans, the copper transporter ATP7A appears to be critical for melanogenesis as well (35). In A. fumigatus, the effect of copper on laccases Abr1 and Abr2 is not known, and the identity of the copper transporter needed for the DHN melanin biosynthesis has yet to be revealed.

In this study, we examined the expression of the laccase-encoding genes abr1 and abr2 during different development stages and corroborated the findings with the temporal and spatial protein expression pattern of Abr1. We found that brlA is essential for turning on the expression of both laccases during conidiation at earlier stages, while abaA and wetA likely play a role in turning off their expression during later stages of conidiation. We also identified a copper transporter gene, ctpA, and demonstrated its importance in the conidial melanization under copper-limiting conditions. We further demonstrated that a melanin-defective cptAΔ mutant, like DHN melanin gene deletion mutants, elicits enhanced immune responses from Galleria mellonella independent of spore viability.

MATERIALS AND METHODS

Strains and media.

A. fumigatus strains, plasmids, and primers used in this study are shown in Tables S1 and S2 in the supplemental material. Strains were grown on Aspergillus complete medium (CM) and minimal medium (MM) (36) or yeast nitrogen base medium (YNB) with appropriate supplements at 30°C in the dark for 3 days as indicated below.

Generation of GFP-tagged abr1 and fluorescence microscopy.

To generate the P_abr1-abr1::eGFP fusion construct, a 5.3-kb fragment that contains 2,879 bp upstream of abr1 along with its full coding sequence except the stop codon was amplified with primers abr1GFPF1 and abr1GFPR1. Similarly, a 2,690-bp GA5-eGFP-AfpyrG fragment was amplified from pFN03 (37) using the primers abr1GFPF2 and GFP Reverse. The two amplified fragments were then fused by overlap PCR to create a 5.4-kb construct, P_abr1-abr1::eGFP, using the primers abr1nesF and GFPnesR. The resulting construct of the fluorescence fusion Abr1 protein under the control of its native promoter was then transformed into the abr1Δ mutant to examine its ability to restore conidial pigmentation. The standard protoplasting method was used for transformation as previously described (38). The same P_abr1-abr1::eGFP construct was also used to transform strains CEA17 and Δabr1.2. Transformants were screened for the presence of abr1::eGFP by PCR and microscopic observations. Further, images were acquired and processed with a Zeiss Axioplan 2 imaging system with AxioCam MRm camera software (Carl Zeiss Microscopy). Green fluorescent protein (GFP) was visualized using the filter FL filter set 38 HE EGFP (Carl Zeiss Microscopy). Samples were prepared and microscopic observation was performed as previously described (39).

Identification of ctpA in A. fumigatus.

The A. fumigatus ortholog of the A. nidulans copper transporter ygA was identified by reciprocal BLASTP searches against the Aspergillus genomic database hosted by the Broad Institute (http://www.broadinstitute.org/annotation/genome/aspergillus_group/Blast.html). The identified A. fumigatus homolog, Afu4g12620, was named ctpA (copper transporter). The subcellular localization of the encoded protein was predicted using the WoLF PSORT program (http://www.genscript.com/psort/wolf_psort.html). The transmembrane helices and the protein structure of CtpA were predicted using the Tmpred program (http://embnet.vital-it.ch/software/TMPRED_form.html) and the I-Tasser program (http://zhanglab.ccmb.med.umich.edu/I-TASSER) (40), respectively.

Generation of the ctpAΔ strain.

The ctpAΔ deletion construct was generated using the fusion PCR protocol as previously described (37). Briefly, a 1,137-bp fragment upstream of the ctpA open reading frame (ORF) was amplified by PCR using primers ctpAF1 and ctpAR1. A 1,175-bp fragment downstream of the ctpA open reading frame (ORF) was amplified using primers ctpAF2 and ctpAR2. The 2,232-bp hygromycin marker fragment was amplified from plasmid pBHt2 (41) using the HYG-construct-F and HYG-construct-R2 primer set. These two PCR products of the ctpA ORF flanking sequences were then fused to the hygromycin marker using the nested primers ctpAnesF and ctpAnesR. This resulted in the final ctpA deletion construct of 4,418 bp. The ctpA deletion construct was then transformed into strain CEA17 through the standard protoplasting method as previously described (38). Transformants were selected for resistance to hygromycin. The deletion of the ctpA gene in selected transformants was verified by PCR and restriction enzyme digestions. The single integration event of the deletion construct at the ctpA locus through homologous recombination in the knockout candidates was further verified by Southern blotting analyses (see Fig. S6 in the supplemental material). For Southern blot analyses, genomic DNA from these candidates was isolated using standard methods (42) and then digested with appropriate restriction enzymes. The restriction enzymes selected recognize one digestion site in the deletion construct. The probe was amplified using primers LB-UI and AI77, and the amplified probe spanned the coding sequence of the hph gene. The probe was labeled with ³²P using the Prime-It random primer labeling kit (Stratagene, Santa Clara, CA) according to the manufacturer's instructions.

Complementation of the ctpAΔ mutant.

To complement the defect of the ctpAΔ mutant, we constructed the wild-type copy of the ctpA gene with its native promoter fused with the GFP gene P_ctpA-ctpA::eGFP by fusion PCR. Briefly, a 4.2-kb fragment that contains 675 bp upstream of ctpA along with its full coding sequence except the stop codon was amplified with primers ctpAGFPF1 and ctpAGFPR1. Similarly, a 2,663-bp GA5-eGFP-AfpyrG fragment was amplified from pFN03 (37) using primers ctpAGFPF2 and GFP Reverse. The two amplified fragments were then fused by overlap PCR using the primers ctpAnesF and GFPnesR. The P_ctpA-ctpA::eGFP construct was then transformed into the ctpAΔ mutant. Transformants were tested for the presence of both ΔctpA::hph and ctpA::eGFP and examined for conidial melanization under copper-limiting conditions.

Generation of the ctpA overexpression strains.

To generate the ctpA overexpression construct, the ctpA open reading frame without the stop codon was amplified using the primer set ctpAoveF and ctpAoveR. The forward primer was designed with an overhang that contains an FseI restriction enzyme digestion site. The reverse primer was designed with an overhang that contains a PacI digestion site. The PCR amplicons were digested with PacI and FseI, and the digested products were then ligated into the overexpression vector pBCtef 27 to generate plasmid pBCtefctp. Plasmid pBCtef 27 carries the hygromycin resistance marker and also the promoter region (∼1 kb) and the terminator region (582 bp) of the translation elongation factor A gene (tefA). The P_tefA and the T_tefA regions were fused together through overlap PCR with an spe1 overhang at the junction of the fused product. Plasmid PBCtef 27 was generated by digesting the fused PCR amplicons with spe1 and then ligating it into the vector pBCHygro at the spe1 site.

Phenotypic assays.

For phenotypic characterizations, 1 × 10⁴ conidia of the wild type, the gene deletion mutant, the complemented strain, and the gene overexpression strain were used. The copper-limiting condition was generated by the addition of the copper chelator bathocuproine disulfonate (BCS; 150 μM) to the media, while the copper-replete condition was generated by the addition of CuSO₄ (500 μM) to the media. We also tested the effect of iron by adding bipyrimidine (50 μM), the iron chelator bathophenanthroline disulfonic acid (BPS; 100 μM), and FeCl₃ (500 μM) to the growth media. Strains were also tested for their sensitivity to oxidative stresses using hydrogen peroxide (6 mM) and NaNO₂ (25 mM) and to osmotic stress using NaCl (0.6 M), KCl (0.6 M), KOH (25 mM), sucrose (1 M), and sorbitol (1 M). Strains were also tested for their resistance to the antifungal drug caspofungin (10 μg/ml) and the detergent SDS (50 μg/ml).

Gene expression studies using quantitative real-time PCR.

For gene expression studies, experiments were conducted and samples were collected as previously described (39). The wild-type strain and the ctpAΔ mutant were grown in 100 ml of complete liquid medium with 1.2 M sorbitol at 30°C with or without the addition of BCS or CuSO₄. Five-milliliter samples of fungal tissues were collected at 12 h and 24 h after conidial inoculation and were labeled as samples from these time points (Fig. 1C). At 24 h postinoculation, 25-ml samples of the cultures were filtered by vacuum through Whatman filter paper no. 2. The fungal mycelial mats were then transferred to solid media and incubated at 30°C for an additional 12 h or 24 h (Fig. 1C; see also Fig. S1 in the supplemental material). These samples were labeled as samples from the 36-h or 48-h time point, respectively. For the brlAΔ, abaAΔ, and wetAΔ mutant samples, only samples from the 24-h, 36-h, and 48-h time points were collected. These mutant strains were grown for 12 h in complete medium at 30°C, then transferred to complete medium with 1.2 M sorbitol, and incubated for another 12 h. Samples were collected at this time point and labeled as being from the 24-h time point. The rest of the experiment and the collection of the mutant samples from the 36-h and 48-h time points were carried out as described above for the wild type and the ctpAΔ mutant strains. Three biologically independent experiments were performed. Total RNA from collected samples was extracted using an RNA purification kit (Invitrogen, CA) by following the instructions from the manufacturer. The RNA samples were treated with TURBO DNA FREE (Invitrogen) to remove any contamination from genomic DNA. The integrity of the RNA samples was examined by gel electrophoresis at multiple steps. The lack of genomic DNA contamination of DNase-treated RNA samples was verified by PCR. The first-strand cDNA synthesis was carried out using a SuperScript III kit (Invitrogen) according to the manufacturer's instructions. KAPA SYBR FAST qPCR master mix was used for the preparation for real-time PCR, and the reaction was performed in an Eppendorf RealPlex 2 machine according to the manufacturer's instructions (KAPA Biosystems). Gene expression levels were normalized based on the tefA expression level of the same samples. Primers ctpArtpcrF and ctpArtpcrR were used for ctpA, primers crtArtpcrF and crtArtpcrR were used for crtA, primers abr1rtpcrF and abr1rtpcrR were used for abr1, primers abr2rtpcrF and abr2rtpcrR were used for abr2, primers alb1rtpcrF and alb1rtpcrR were used for alb1, primers arp1rtpcrF and arp1rtpcrR were used for arp1, primers arp2rtpcrF and arp2rtpcrR were used for arp2, and primers tefrtpcrF and tefrtpcrR were used for tefA. The sequences of these primers are included in Table S2 in the supplemental material.

Statistical analyses.

Statistical significance of gene expression among different groups was assessed by one-way analysis of variance (ANOVA) nonparametric tests. Pairwise two-tailed t test nonparametric analysis was used to analyze the differences between two groups. All statistical analyses were performed using the Graphpad Prism 5 program, with P values lower than 0.05 considered statistically significant.

Inoculation of G. mellonella larvae with A. fumigatus conidia.

Larvae of the great wax moth, G. mellonella, in the final instar larval stage were obtained from Vanderhorst, Inc. (St. Marys, OH). Larvae of about 0.3 to 0.4 g were used for inoculation as previously described (43). Briefly, 1 × 10⁶ A. fumigatus conidia in 5 μl of phosphate-buffered saline (PBS) were injected into the hemocoel of each wax moth via the last left proleg. After injection, the larvae were incubated in plastic containers at 37°C and photographed at various time points. For each experiment, 10 to 15 larvae per group were infected. All experiments were replicated at least twice. Heat-killed conidia were prepared by heating the conidia in PBS suspension at 100°C for 10 min. UV-killed conidia were prepared by irradiating 1 × 10⁸ conidia in 10 ml of water in petri dishes for an hour using the default time mode function of an XL-1500 UV cross-linker with an intensity of ∼6,000 μW/cm² (Spectronics Corporation, NY). The killing of the conidia by the heat or the UV treatment was confirmed by plating treated conidia on complete medium and culturing them for an additional 3 days at 30°C. The conidial suspension was then washed twice with PBS prior to injection. Larvae were infected with live conidia, UV-killed conidia, heat-killed conidia, or mixtures of live and dead spores of different combinations as indicated below. The mixtures were prepared using equal numbers of dead spores and live spores of indicated strains.

The hemolymph of G. mellonella was collected as previously described (44), with minor modifications. Briefly, G. mellonella larvae were made to bleed by making a small incision on one of the prolegs. A fresh hemolymph drop was then squeezed out by applying gentle pressure and collected directly onto a glass slide containing 10 to 20 μl of anticoagulant solution (100 mM sodium phosphate buffer, pH 7). A glass coverslip was overlaid gently on top of the hemolymph. Before putting coverslips on top of the hemolymph, spacers (made from coverslips) were placed on both sides of the hemolymph to reduce any turbulence to hemocytes by the overlaid coverslips. These freshly prepared slides were then used immediately for microscopic examination.

RESULTS

The expression of laccase genes abr1 and abr2 and other melanin genes is developmentally regulated and is controlled by master regulators of conidiation.

To facilitate the examination of gene expression at different developmental stages, we classified the growth cycle of A. fumigatus as follows. Dormant conidia break dormancy and germinate when environmental conditions become favorable. Germinated conidia then switch to polar growth to produce elongated vegetative hyphal cells, which develop further into interconnected mycelia that are competent for conidiation (Fig. 1A). Upon exposure to air, competent hyphae produce aerial stalks to initiate the development of conidiophores with elaborate multicellular structures that subsequently generate chains of uninucleate conidia (Fig. 1A). The conidia become heavily pigmented during maturation due to increased deposition of the DHN melanin (Fig. 1B) (21, 45, 46). To examine the expression pattern of abr1 and abr2 genes, we collected wild-type A. fumigatus samples at 12 h (vegetative hyphae), 24 h (competent hyphae), 36 h (early conidiophores), and 48 h (conidiophores with mature conidia) (Fig. 1C; see also Fig. S1 in the supplemental material). As expected, the expression of both the abr1 and abr2 genes was low during vegetative hyphal growth, but it was dramatically increased during conidiation (Fig. 2A). We further examined the expression pattern of other genes from the melanin gene cluster (alb1, arp1, and arp2) and found that their expression was also increased during conidiation (Fig. 2B). The expression dynamics of the abr1, abr2, alb1, arp1, and arp2 genes during development were consistent with previous studies showing that the DHN melanin gene cluster is highly expressed during conidiation (22, 28, 29).

Fig 2.

The expression of abr1 and abr2 and other melanin genes is developmentally regulated and subject to the control of master developmental regulators. (A) Expression of alb1, arp1, arp2, abr1, and abr2 during development. All gene expression levels were normalized based on the tefA expression levels. The 12-h time point was used as the reference for the subsequent time points. (B) Melanin gene expression levels in the wild type and the brlAΔ, abaAΔ, and wetAΔ mutants. All gene expression levels were normalized based on the tefA expression levels. The expression level of each gene at 24 h was used as the reference for the subsequent time points (fold changes with standard deviations are shown).

To assess if the gene expression pattern correlates with the protein expression pattern, we constructed GFP-fused Abr1 controlled by the abr1 native promoter (P_abr1-Abr1-GFP). The P_abr1-Abr1-GFP construct, when introduced into the brownish abr1Δ mutant, was able to restore the mutant conidial pigmentation to the wild-type color (see Fig. S2 in the supplemental material), indicating the functionality of this fluorescence-tagged Abr1. Consistent with the gene expression pattern, the Abr1-GFP fluorescence signal was not detected in vegetative hyphae, but it was detected in aerial hyphae, vesicles, phialides, and young conidia (Fig. 3). The fluorescence signal became stronger in mature aerial stalks, vesicles, phialides, and conidia than that in young aerial stalks. The distribution of Abr1-GFP was not even along the cell surface and was more concentrated at the apical region in the newly generated aerial stalks (Fig. 3m), at the area of vesicles where phialides would emerge (Fig. 3g and h), and at the apex of phialides where conidia would appear. This protein was found as small bright patches decorating along the cell surface of stalks, vesicles, and conidia (Fig. 3m, n, o, and p). The fluorescence-tagged Abr1 displayed similar expression and subcellular localization patterns when it was induced to the abr1Δ mutant or the wild type.

Fig 3.

Localization of Abr1-GFP. Abr1-GFP is observed on the cell surface of aerial hyphae, vesicles, phialides, and conidia. No GFP is observed in vegetative hyphae. (a and b) Vegetative hyphae; (c and d) aerial hyphae; (e and f) young vesicles; (g and h) bright spots observed in the area of vesicles from which phialides would emerge; (i and j) vesicles with phialides; (k and l) conidia. (m and n) Abr1-GFP was observed as small patches (arrows) decorating the lines of the cell surface and the tip of the aerial stalk. Scale bars (a to n), 10 μm. (o and p) In conidia, Abr1-GFP is observed at the cell surface. The white arrow shows small patches. Scale bar, 5 μm.

The induction of the expression of the laccase genes commenced at hyphal competency that precedes the conidiation event, and their expression reached exceptionally high levels during conidiation (Fig. 2A and 3). A similar expression pattern was also observed for all other tested genes, alb1, arp1, and arp2, from the DHN melanin gene cluster (Fig. 2B). This led us to hypothesize that master regulators of Aspergillus development are likely involved in their regulation. The genes brlA, abaA, and wetA are known to be required for early, middle, and late conidiation, respectively, in Aspergillus (5). The deletion of brlA blocks Aspergillus development at the stage of vesicle development prior to the formation of the conidiophore (4, 5), abaA is required for the generation of conidia from phialides (analogous to stem cells for conidia), and wetA is important for conidial maturation (6, 34). Therefore, we examined the impact of the disruption of these regulators on the expression of abr1 and abr2, alb1, and arp1 and arp2. The deletion of the brlA gene abolished the induction of the expression of abr1 and abr2, alb1, and arp1 and arp2. The expression levels of these genes in the brlAΔ mutant remained low at all the time points examined (Fig. 2A and B). This result is in accord with the essential role of BrlA in hyphal competency and conidiation (2, 5). Surprisingly, the expression pattern of the melanin genes was drastically different in the abaAΔ and the wetAΔ mutants. Specifically, the expression levels of abr1 and abr2 and all other melanin genes tested (alb1, arp1, and arp2) at 24 h (hyphal competency) and 36 h (early conidiophore development) were exceedingly high in the abaAΔ and wetAΔ mutants (Fig. 2B). At 48 h, the expression levels of these melanin genes in the abaAΔ and wetAΔ mutants were more or less at high levels comparable to those observed in the wild type at the stage of conidiophore maturation. The high levels of expression observed for these melanin genes are consistent with the fact that colonies of the abaAΔ and wetAΔ mutants are still melanized, while the brlAΔ mutant is not (see Fig. S3 in the supplemental material). These results suggest that abaA and wetA are not necessary for the induction of the expression of abr1 and abr2 and other melanin genes. Rather, abaA and wetA are necessary for the suppression of their expression at earlier stages of development prior to conidiophore maturation. Given the importance of these central regulators in development, it is not clear if these regulators control the expression of these melanin genes directly or indirectly due to their impact on fungal development.

The induction of the expression of Abr1 and Abr2 at hyphal competency is responsive to copper limitation, and CtpA is a developmentally regulated putative copper transporter.

Copper has been suggested to stimulate laccase gene expression in C. neoformans (47, 48). To examine if copper regulates the expression of the laccase genes abr1 and abr2 in A. fumigatus, we examined the expression pattern of abr1 and abr2 under copper-rich and copper-limiting conditions. The expression pattern of abr1 and abr2 was not affected by excessive copper present in the growth media (Fig. 4A and B). However, under the copper-limiting condition, we observed modestly enhanced induction of abr1 and abr2 expression during the hyphal competent stage (Fig. 4A and B). Thus, copper starvation appears to augment the induction of the expression of abr1 and abr2. The presence or the lack of copper in the media did not improve or exacerbate the pigmentation defects in the mutants in which the melanin genes had been deleted (Fig. 4E).

Fig 4.

Identification of a developmentally regulated copper transporter, ctpA. (A and B) The induction of the expression of abr1 and abr2 is significantly enhanced at hyphal competency stage as well as late conidiation stage by copper starvation in the wild type. The P value for the abr1 expression level among all the groups at the 24-h time point was 0.0059, indicating statistical significance. The P values for the pairwise comparisons of abr1 expression for the 24-h time point were 0.1422 (CM and CuSO₄), 0.0383 (CM and BCS), and 0.0255 (CM and ctpAΔ). The P value for the abr1 expression level among all the groups at the 36-h time point was 0.4701 and was not considered significantly different. The P value for the abr1 expression level among all the groups at the 48-h time point was 0.0002, indicating statistical significance. The P values for the pairwise comparisons for the 48-h time point were 0.8892 (CM and CuSO₄), 0.0035 (CM and BCS), and 0.0424 (CM and ctpAΔ). Similarly, the P value for the abr2 expression level among all the groups at the 24-h time point was 0.0145, indicating statistical significance. The P values for the pairwise comparisons of abr2 expression for the 24-h time point were 0.0899 (CM and CuSO₄), 0.0922 (CM and BCS), and 0.0293 (CM and ctpAΔ). The P value for the abr2 expression level among all the groups at the 36-h time point was 0.4692 and was not considered significantly different. The P value for the abr2 expression level among all the groups at the 48-h time point was 0.0071, indicating statistical significance. The P values for the pairwise comparisons for the 48-h time point were 0.9942 (CM and CuSO₄), 0.0104 (CM and BCS), and 0.1043 (CM and ctpAΔ). Thus, the expression of both abr1 and abr2 is significantly enhanced at the hyphal competency stage in the ctpAΔ mutant compared to the wild type and also in the wild type under copper starvation conditions compared to that under normal conditions. The ctpAΔ mutant was grown in CM. RNA extracted from fungal cultures at indicated developmental stages was used to measure gene expression levels. All gene expression levels were normalized based on the tefA expression levels. The 12-h time point was used as the reference for the subsequent time points. (C) Expression pattern of ctpA in the wild-type A. fumigatus during development and in media with varied levels of copper. The expression of ctpA is enhanced significantly only compared to that under the CuSO₄ andBCS conditions at 24 h (P value, 0.0080). WT, wild type; CuSO₄, CM with addition of copper; BCS, CM with addition of the copper chelator. The 12-h time point was used as the reference for the subsequent time points. (D) Expression pattern of ctpA in the wild type and the brlAΔ, abaAΔ, and wetAΔ mutants. The ctpA expression pattern was statistically significantly different between the wild type and the brlAΔ mutant only at the 48-h time point (P value, 0.0160). The 24-h time point was used as the reference for the subsequent time points. Error bars indicate standard deviations. (E) The strains were cultured on CM and media with CuSO₄ or BCS at 30°C for 2 days.

To identify the copper transporter that is likely involved in conidial melanization in A. fumigatus, we performed reciprocal BLASTP searches of the A. fumigatus genome using A. nidulans ygA and C. lindemuthianum clap1 as the queries. These copper transporters are shown to be involved in melanization (32, 34). The gene Afu4g12620 was identified as the homolog that shares the highest similarity with the query sequences, and we named it ctpA. The ORF of this gene is 3,564 bp, and it is predicted to encode a protein of 1,167 amino acids (see Fig. S4 in the supplemental material). The amino acid sequence of CtpA is 72% identical and 82% similar to that of A. nidulans YgA, and it shares 76% similarity to C. lindemuthianum Clap1. All the signature motifs found in the Clap1 and YgA, including the copper ion binding domains defined by the sequence GMTCGAC and ATP-binding sites, are conserved in CtpA and are present in the same order (see Fig. S4 and S5A in the supplemental material). Furthermore, CtpA likely carries 8 transmembrane helices, as predicted by the Tmpred program and the I-Tasser program (see Fig. S4 and S5), and consistently, CtpA is predicted to be a membrane protein by the WoLF PSORT program. The best-matched structural homolog from the protein database is 3rfuA, which is a copper transporting PIB-type ATPase (see Fig. S5B).

We speculate that if ctpA is involved in DHN melanization in conidia, its activity might be upregulated during hyphal competency and/or conidiation. Indeed, the ctpA expression was low during vegetative hyphal growth, but it peaked at the hyphal competency developmental stage (Fig. 4C). ctpA expression then decreased after hyphal competency, although it was still maintained at a higher level during conidiation than during vegetative hyphal growth. Surprisingly, the addition of copper or the copper chelator BCS to the growth media did not significantly alter ctpA expression (Fig. 4C), suggesting that ctpA is not regulated by extracellular copper at the transcript level. As the expression of ctpA is developmentally regulated, we decided to examine if the master development regulators control its expression. We analyzed the expression of ctpA in the brlAΔ, abaAΔ, and wetAΔ mutants and compared it to that of the wild-type control. As shown in Fig. 4D, ctpA expression failed to diminish during conidiation in the brlAΔ mutant, suggesting that BrlA is involved in regulating the expression of ctpA after hyphal competency. The effect of the deletion of abaA or wetA on ctpA expression was not conclusive because of less obvious effect and large variations.

Disruption of ctpA leads to defective conidial pigmentation that can be remediated by introducing a wild-type copy of ctpA or by the addition of copper.

To examine the possible role of CtpA in conidial melanization, we deleted the ctpA gene in A. fumigatus. The replacement of the ctpA gene with the hygromycin marker through homologous replacement was verified by Southern blotting (see Fig. S4C to E in the supplemental material). We also generated a ctpA overexpression (ctpA^oe) strain by placing the wild-type copy of the ctpA gene under the control of the promoter of the housekeeping gene tefA. We selected the tefA promoter based on our observation that tefA was constitutively expressed at high levels in this fungus throughout its life cycle (data not shown).

Although the wild-type strain, the ctpAΔ mutant, and the ctpA^oe strain displayed normal vegetative hyphal growth on a variety of culture media, the ctpAΔ mutant showed a severe defect in conidial melanization under certain conditions. For instance, the ctpAΔ mutant grown on typical growth media was poorly melanized, resembling the wild-type strain grown under extremely copper-limiting conditions (Fig. 5A and B). Addition of copper to the growth media restored the conidial pigmentation of the ctpAΔ mutant to the wild-type color (Fig. 4E and 5A and B).

Fig 5.

ctpA is critical for conidial melanization under normal and copper-limiting growth conditions. (A) The wild type, the ctpAΔ mutant, and the ctpA^oe strain were grown in complete sorbitol medium, with or without CuSO₄ or BCS. (B) The wild type, the ctpAΔ mutant, and the ctpA^oe strain were grown in YNB sorbitol medium, with or without CuSO₄ or BCS. The different media have different levels of copper availability. (C) The ctpAΔ mutant and the complemented strain (ctpAΔ plus P_ctpA-CtpA-GFP) were grown in complete medium, with or without CuSO₄ or BCS.

To further confirm that the defect in conidial pigmentation observed in the ctpAΔ mutant was indeed caused by the lack of the ctpA gene, we constructed GFP-fused CtpA controlled by the ctpA native promoter (P_ctpA-CtpA-eGFP). The P_ctpA-CtpA-GFP construct, when introduced into the ctpAΔ mutant, was able to confer the wild-type conidial pigmentation on the mutant under both the typical and the copper-limiting conditions (Fig. 5C). Thus, ctpA is required for conidial pigmentation under normal and copper-limiting conditions.

CtpA-GFP fluorescence was detected in all development stages, with a slightly weaker signal in vegetative hyphae and a stronger signal in conidiophores (Fig. 6). The fluorescence signal of CtpA-GFP has tubular or punctate shapes, resembling secretion vesicles. CtpA protein expression appears to follow development, as fluorescence was brighter in actively growing cells, like vesicles and phialides in young conidiophores. In more mature conidiophores, fluorescence in vesicles became dimmer but remained bright in phialides (stem cells) and conidia (Fig. 6).

Fig 6.

Localization of CtpA-GFP. Fluorescence of CtpA-GFP was detected in the vegetative hyphae (a and b), young conidiophore (c and d), mature conidiophore with conidia (e and f), and conidia (g and h). CtpA fluorescence signals take shapes resembling secretion vesicles. Scale bars, 10 μm (a to f) and 5 μm (g to h).

Previous studies have implied interconnection between copper and iron homeostasis (49, 50). Hence, we also tested conidial melanization of the ctpAΔ mutant in response to iron limitation. The ctpAΔ mutant was much more severely impaired in conidial pigmentation when grown on media that contained bipyrimidine, a chemical that has a high affinity for iron but also chelates copper (51) (Fig. 7). Since the effect of bipyrimidine on pigmentation of the ctpAΔ mutant could be due to a lower copper level, we decided to test the effect of the iron chelator bathophenanthroline disulfonic acid (BPS), which has been used to specifically study iron regulation in a variety of organisms, including another fungal pathogen, Cryptococcus neoformans (52). As shown in Fig. 7, the ctpAΔ mutant pigmented at a reduced but similar level when grown on media with or without BPS, indicating that BPS, or iron limitation, does not as significantly affect conidial pigmentation of the ctpAΔ mutant as copper limitation. Consistently, the addition of FeCl₃ to the growth media failed to significantly improve ctpAΔ mutant pigmentation (Fig. 7). In contrast, the addition of CuSO₄ to the growth media fully restored the melanin defects of the ctpAΔ mutant (Fig. 5). Taken together, these results suggest that ctpA encodes a copper-specific transporter involved in melanization.

Fig 7.

The conidial pigmentation in the ctpAΔ mutant is not significantly affected by iron limitation or excess in the medium. The wild type, the ctpAΔ mutant, and the ctpA^oe strain were grown in CM with bipyrimidine, the iron chelator BPS, and FeCl₃. The wild-type strain produced melanized conidia under these conditions. The ctpAΔ mutant showed a severe defect in conidial pigmentation when grown in bipyrimidine, but its pigmentation was not significantly affected by BPS or FeCl₃.

The expression pattern of abr1 and abr2 in the ctpAΔ mutant resembles that under copper-limiting growth conditions.

As copper limitation modestly enhances the expression of abr1/2 at the hyphal competency stage (Fig. 3A and B), we speculated that the disruption of the copper transporter CtpA would resemble copper starvation and may similarly affect the expression of these laccase genes. Indeed, the induction of the expression of both abr1 and abr2 in the ctpAΔ mutant strain grown under normal conditions was significantly enhanced at the stage of hyphal competency, mimicking the pattern observed in the wild-type strain grown under copper-limiting conditions (Fig. 4A and B). Thus, the lack of ctpA in A. fumigatus creates a response similar to the one toward copper starvation. This finding is consistent with the poor conidial melanization of the wild-type A. fumigatus under copper-limiting conditions and of the ctpAΔ mutant under both normal and copper-limiting growth conditions (Fig. 5B).

Pigment-defective conidia of the ctpAΔ mutant stimulate host immune responses in the Galleria mellonella infection model.

We showed previously that conidia of all the DHN melanin biosynthesis mutants, including abr1 and abr2 mutants, cause rapid darkening of infected G. mellonella larvae (53). Such darkening is most likely due to the formation of melanotic capsules by the host, a key immune defense mechanism in insects (44, 54). We hypothesize that the lack of melanin in these mutants exposes immunogenic molecules on the A. fumigatus cell wall and consequently elicits strong immune responses from this insect (53). As the copper transporter ctpAΔ mutant exhibited a provisional defect in conidial melanization under normal and copper-limiting growth conditions (Fig. 5), we speculated that pigment-defective ctpAΔ conidia collected under such growth conditions would elicit immune responses from the larvae as strong as those induced by the DHN melanin mutants. Indeed, G. mellonella larvae rapidly darkened after being inoculated with pigment-defective conidia of the ctpAΔ mutant (Fig. 8A). In contrast, the normal-pigmented conidia of the ctpAΔ mutant collected under copper-excessive growth conditions did not cause increased darkening in the infected larvae compared to those inoculated with the wild-type pigmented conidia (Fig. 8A). The pigmented ctpAΔ mutant conidia and wild-type conidia both induced only modest darkening in the larvae compared to that in the PBS-inoculated larvae (Fig. 8A).

Fig 8.

Melanin-defective ctpAΔ mutant stimulates G. mellonella host immune responses. (A) The A. fumigatus wild-type strain and the ctpAΔ mutant were grown on CM with or without the addition of BCS or CuSO₄. Conidia of the strains grown under these conditions were collected and used for larva inoculation. Larvae injected with PBS were used as a control. Images shown here are photos of the inoculated larvae taken at 16 h after inoculation. Darkening of the larvae was apparent within 4 h (data not shown). Larvae responded to the ctpA^oe conidia similarly to the wild-type conidia and thus are not shown here. (B) Heat-killed conidia of the ctpAΔ mutant harvested from growth media containing CuSO₄ or BCS were used to inoculate G. mellonella larvae. A mixture of heat-killed conidia of the ctpAΔ mutant harvested from CM-BCS media and live conidia of the ctpAΔ mutant harvested from CM plus CuSO₄ showed modest darkening of the larvae. Reversed combination yielded similar results, which are not shown here. Combination of the ctpAΔ mutant and the wild-type conidia also yielded similar results, not shown here. Larvae injected with the PBS buffer were used as a control.

Microscopic observation of the hemolymph (equivalent of blood in higher animals) isolated from larvae infected by the pigment-defective ctpAΔ conidia and those infected by the normal-pigmented ctpAΔ conidia revealed differences in melanized hemocytic aggregates (Fig. 9). Melanization accompanying phagocytosis and hemocyte aggregation (melanotic capsules) is one main means for G. mellonella and other insects to contain foreign invaders (44, 54). Much larger melanotic capsules were observed in the hemolymph isolated from larvae challenged with the pigment-defective conidia harvested from the ctpAΔ mutant grown under copper-limiting conditions than those on copper-rich media (Fig. 9). No difference was observed between pigmented ctpAΔ mutant conidia and pigmented wild-type conidia in eliciting the formation of melanotic capsules in larvae.

Fig 9.

G. mellonella exhibits earlier and stronger hemocyte-mediated defense responses when infected with pigment-defective ctpAΔ conidia. Images are of G. mellonella larvae taken at 16 h after inoculation with 1.0 × 10⁶ ctpAΔ mutant conidia harvested from copper-limiting or copper-rich conditions. Larger capsules produced by the larva hemolymph (aggregations composed of a variety of hemocytes to contain foreign particles) were visible when inoculated with pigment-defective ctpAΔ conidia harvested under copper-limiting conditions (BCS) compared to those infected by the pigmented ctpAΔ conidia harvested under copper-rich conditions. Some of the capsules were heavily melanized (melanotic capsule) by the larvae. Larvae injected with the PBS buffer were used as a control. Scale bar, 20 μm.

If the rapid and enhanced immune responses of larvae toward the melanin-defective mutants are caused by structural changes in the conidial cell wall, such responses should be independent of conidial viability. To test this idea, we heat killed the conidia of the ctpAΔ mutant harvested under different growth conditions and used them for the inoculation. Similar to what we observed with live conidia, the heat-killed pigment-defective conidia of the ctpAΔ mutant caused severe darkening in the larva cuticle (Fig. 8). We carried out similar experiments using UV-killed pigment-defective conidia of the ctpAΔ mutant, and the results were similar to those for the heat-killed conidia (see Fig. S6 in the supplemental material). In contrast, heat-killed pigmented conidia of the ctpAΔ mutant did not elicit enhanced larva darkening or the formation of melanotic capsules compared to the wild-type conidia (Fig. 9; see also Fig. S7 in the supplemental material). When a mixture of dead pigment-defective ctpAΔ conidia and live pigmented ctpAΔ conidia or vice versa was used to inoculate larvae, a modest enhancement in larva darkening was observed (Fig. 7B). Furthermore, coinoculation of live wild-type conidia with heat-killed pigment-defective ctpAΔ conidia modestly increased larva darkening compared to the live wild-type conidia, with the same total fungal cell number (dead and live) (data not shown). These results suggest that an active metabolism or active secretion of proteins or small molecules exhibited by viable cells is not responsible for this enhanced larva immune response. Rather, alterations in the conidial surface of the pigment-defective ctpAΔ mutant might be responsible. This is in accordance with our previous finding that rapid and intense cuticle darkening of G. mellonella larvae in response to the DHN melanin mutants is independent of conidial viability. In this regard, the ctpAΔ mutant grown under copper-limiting conditions behaves similarly to the DHN melanin mutants.

DISCUSSION

Melanin biosynthesis is one of the most universal phenomena, observed in a variety of life forms, including plants, animals, bacteria, and fungi. However, the regulation of melanin production is uniquely adopted by living organisms in their adaptation to variable environmental conditions and their lifestyles (55). Not surprisingly, melanin is often synthesized in a cell-type-specific manner (56): Melanin is deposited in the appressorium to provide rigidity and selective permeability to its cell wall, and it is required for pathogenicity in the plant pathogens Colletotrichum orbiculare and Magnaporthe oryzae (57, 58). The exclusion of melanin at the penetration site implies a role in determination of the correct invasion point (59). In this study, we focused on abr1 and abr2 in the A. fumigatus DHN melanin cluster, as laccases are the most conserved factors required for melanin production irrespective of the type of melanin or the organism involved. We found that the expression of both laccase genes together with other melanin genes was induced upon hyphal competency and was maintained at even higher levels during conidiation (Fig. 2). Using fluorescence-tagged Abr1, we confirmed that this laccase was in stalks and conidiophore structures (Fig. 3). The fluorescent Abr1 mostly delineated the cell periphery as small bright patches, consistent with the deposition of melanin granules in the cell wall.

In Colletotrichum, melanization in mycelia and appressoria is controlled by different transcription factors (60). However, the A. fumigatus melanin cluster does not carry any regulatory factor (10, 22), unlike some other fungi (61). As A. fumigatus produces hyaline vegetative hyphae and pigmented conidiophores, it is logical to reason that DHN melanin genes are upregulated during conidiation, and the key development regulators might control their expression. BrlA, AbaA, and WetA are the central development regulators in Aspergillus spp. What is different in A. fumigatus compared to its model relative A. nidulans is that the expression of abaA and wetA is not completely dependent on brlA (28). In this study, we found that BrlA, AbaA, and WetA exert different impacts on the expression of abr1 and abr2 and other melanin genes (Fig. 2B): BrlA is required for the induction of their expression at the hyphal competency, while AbaA and WetA appear to act to suppress the expression of these melanin genes prior to conidiophore maturation (Fig. 2B). A previous study also indicated a requirement for developmental modifier StuA in regulating the melanin gene cluster (28). Another developmental modifier, MedA, when disrupted, causes a delay in conidial melanization and a modest reduction in melanin gene expression independent of brlA, abaA, or wetA in A. fumigatus (62). Thus, the regulation of melanin genes in A. fumigatus is controlled by multiple developmental regulators in a complicated manner.

Laccases require copper as their cofactor, and dysregulation of copper homeostasis could affect their activity and/or their expression. When laccases are deleted, however, the pigmentation defects cannot be exacerbated by copper limitation or improved by copper repletion (Fig. 4E). In the wild-type strain, however, copper starvation augmented the induction of expression of abr1 and abr2 during hyphal competency (Fig. 4A and B). Interestingly, the disruption of the copper transporter ctpA enhanced the induction of abr1 and abr2 during hyphal competency, mimicking the effect of copper starvation (Fig. 4A and B). The expression of ctpA itself peaked at hyphal competency stage, which is consistent with its potential role in assisting the supply of copper to laccases Abr1/2. Consistent with its predicted role as the copper transporter involved in conidial melanization, ctpA is required for normal pigmentation under normal and copper-limiting conditions in A. fumigatus (Fig. 4E and Fig. 5). The observations that the defect in conidial pigmentation of the ctpAΔ mutant can be caused by copper but not iron limitation and that excessive copper but not iron in the media restored the melanin defect (Fig. 5) suggest that this transporter is a major factor responsible for supplying copper for these laccases. These findings also underscore the significance of the hyphal competency stage in the preparation for conidiation in A. fumigatus. The disruption of ctpA enhanced A. fumigatus resistance toward oxidative stress caused by hydrogen peroxide, but the mutation did not appear to affect its tolerance to other stresses, such as salts (KCl and NaCl), sucrose, SDS, antifungal caspofungin (inhibits β-glucan biosynthesis), or NaNO₂ (data not shown). Thus, it appears that ctpA exerts its primary effect on conidial pigmentation.

Wild-type A. fumigatus produces echinulate conidia and is weakly immunogenic. The melanin mutants have profound modification of the cell wall and display a smooth conidial surface (10, 63). As conidial surface-associated molecules represent the interface between this fungus and its host, it is not surprising that melanin plays multiple roles in modulating A. fumigatus interaction with its host (27, 53, 64). Here we showed that both dead and live pigment-defective ctpAΔ conidia caused rapid darkening of Galleria larvae after inoculation due to drastically enhanced formation of melanotic capsules by the larva hemocytes (Fig. 8 and 9). This reflects highly stimulated host innate immune responses in this insect, consistent with enhanced immune responses in mammalian cells elicited by melanin mutants. A very recent study has shown that melanin mutant conidia are more rapidly cleared from lungs of mice than are wild-type conidia (65). The responses of larvae to the pigment-defective ctpAΔ conidia are the same as what we observed when Galleria larvae were inoculated with the DHN melanin mutants (53). As insects like the great wax moth provide several experimental advantages over mammal models (e.g., reduced cost, easier handling, and fewer bioethics considerations), they have become increasingly popular in assessing host innate immune responses and fungal virulence factors (43, 66–68), including melanin biosynthesis factors in Cryptococcus and Aspergillus. Collectively, our findings and those of others indicate complicated regulation of melanin in A. fumigatus by developmental factors and copper homeostasis and the importance of melanin in shaping the interaction between fungi and various hosts.