LaeA, a Regulator of Morphogenetic Fungal Virulence Factors

Opportunistic animal and plant pathogens, well represented bythe genus Aspergillus, have evolved unique mechanisms to adaptto and avoid host defenses. Aspergillus fumigatus, an increasinglyserious pathogen owing to expanding numbers of immunocompromisedpatients, causes the majority of human infections; however,an inability to identify bona fide virulence factors has impededtherapeutic advances. We show that an A. fumigatus mutationin a developmentally expressed transcriptional regulator ( ${Delta}$ laeA)coordinating morphological and chemical differentiation reducesvirulence in a murine model; impaired virulence is associatedwith decreased levels of pulmonary gliotoxin and multiple changesin conidial and hyphal susceptibility to host phagocytes exvivo. LaeA, a conserved protein in filamentous fungi, is a developmentalregulator of virulence genes and, possibly, the first antimicrobialtarget specific to filamentous fungi that are pathogenic toplants and animals.

INTRODUCTION

Top

Introduction

For opportunistic pathogens, host infection may not be necessary,or even advantageous, for species survival. True "virulence"factors are typically missing in these organisms; instead, speciesthat have adapted mechanisms to persist in hostile environments(including in vivo) become selected as the most common causesof disease. Identification of specific products or pathwayscritical to in vivo adaptation would potentially enable developmentof the most-targeted therapeutics, but the discovery of single-genedisease determinants in opportunists is hampered by redundancyboth in the organism and in the host's response to infection.This is one factor impeding development of effective treatmentsfor opportunistic fungal plant and animal pathogens; in thiscase, the delay has been particularly detrimental because drugstargeting fungal cell wall and fungal membrane components confertoxicities to eukaryotic hosts.

All pathogenic fungi that sporulate or form fruiting bodieshave developed mechanisms to mediate coordinate expression ofbiologically active secondary metabolites and spore-relatedproducts during development (7). The hydrophobins, conidialcell wall proteins that have been adopted by different plantand animal pathogens to serve diverse functions necessary forfungal pathogenicity, are an example. Despite the diversityof amino acid sequence and function, hydrophobins from disparatefungi complement each other in a variety of environmental andpathological niches (16). Thus, mechanisms of developmentalregulation which more broadly control fungal pathogenicity mayprove to be the pivotal virulence factors and drug targets forfilamentous fungi that cause disease in plants and animals.

Members of the genus Aspergillus are common causes of diseasein plants and animals worldwide. Currently, invasive aspergillosisis a leading cause of infection-related mortality in immunocompromisedpatients, occurring in people with hematological malignanciesand AIDS, and limiting the success of hematopoietic stem celland solid-organ transplantation (8, 23, 24). Mortality is highand in the setting of disseminated infection approaches 100%even with the use of newer antifungal agents (10, 20). Aspergillusspecies cause additional environmental concern and morbidityin nonimmunosuppressed people and are implicated as a causeof fungal sinusitis, asthma, and allergic bronchopulmonary aspergillosis(17). Plant-pathogenic aspergilli are infamous for their elaborationof several mycotoxins, including aflatoxin, the most potentnatural carcinogen known to date (38).

Despite the multitude of Aspergillus species that are ubiquitousin our environment, one species, Aspergillus fumigatus, accountsfor at least 90% of invasive infections by Aspergillus spp.in people (5, 17). The clinical significance of this speciesjustified prioritization for genome sequencing, which was recentlycompleted (www.tigr.org). However, few null mutants have demonstratedavirulent phenotypes. For example, single- and multiple-genedisruption in a pigment biosynthesis gene cluster has been associatedwith decreased virulence in mouse models without inducing aconcomitant alteration in microbial fitness; this defect isassociated with increased microbial susceptibility to oxidativekilling by monocytes and polymorphonuclear neutrophils (PMNs)(13, 34, 37).

Recently, we identified a nuclear protein, LaeA, that acts asa regulator of secondary metabolism in the aspergilli, includingthe model organism Aspergillus nidulans, the industrial agentAspergillus terreus, and A. fumigatus (3). Here we show thatthe loss of laeA ( ${Delta}$ laeA) in A. fumigatus, while not producinggross changes in growth or sporulation in media, results inreduced virulence in a murine model. Impairment is associatedwith the loss of detectable gliotoxin, an immunotoxin putativelyinvolved in virulence (18, 30), increased conidial susceptibilityto macrophage phagocytosis, and decreased ability of hyphaeto kill neutrophil cells.

MATERIALS AND METHODS

Top

Materials and Methods

Media and reagents. All fungal strains (Table 1) were maintained as glycerol stocksand were cultured at 25°C or 37°C on glucose minimalmedium (GMM) with appropriate supplements (32), Sabouraud'sdextrose agar (Becton Dickinson, Sparks, MD), or RPMI 1640 (plus10 mM HEPES, pH 7.4, without phenol red; Sigma, St. Louis, MO).Acetoxymethyl ester of calcein (calcein AM), calcofluor white,and fluorescein isothiocyanate (FITC) were purchased from MolecularProbes (Eugene, OR). The complete medium used for macrophageculture included Iscove's modified Dulbecco's medium (Gibco,Grand Island, NY) supplemented with 10% heat-inactivated humanAB serum, 0.3 mg/ml L-glutamine, and 200 U/ml penicillin-streptomycin.Recombinant macrophage colony-stimulating factor was obtainedfrom R&D systems (Minneapolis, MN).

View this table:
[in this window]
[in a new window]
TABLE 1. Strains used in this study

Fungal strains and nucleic acid analysis. The strains used are listed in Table 1. Aspergillus fumigatuslaeA was previously cloned, sequenced, and disrupted in wild-type(WT) strain AF293.1 (a pyrG auxotroph) by the replacement oflaeA with pyrG (3). Briefly, the disruption vector pJW58 wasconstructed by flanking the Aspergillus parasiticus pyrG markergene with DNA fragments from upstream of the A. fumigatus laeAstart codon (0.9 kb) and downstream of the laeA stop codon (1.0kb). Fungal protoplasts were transformed as previously described(3). Transformant TW54.2 was documented to have single-genereplacement of laeA by Southern analysis and PCR. pJW72.1, containinga wild-type copy of the 3-kb laeA gene, including a 1-kb promoter,was used to complement the ${Delta}$ laeA strain TJW54.2 to make TJW68.6and TJW69.2. The 3-kb laeA gene was amplified by primers FUM1and FUM4 (3) and was inserted in the Zero Blunt TOPO vector(Invitrogen) to produce pJW57.2. pJW72.1 was created by ligatingthe 3-kb NotI-SpeI laeA fragment from pJW57.2 into the NotI-SpeIsite of pUCH2-8 (1). pUCH2-8 contains the gene encoding hygromycinB phosphotransferase, which was the selectable marker used fortransformation of Aspergillus.

The plasmid pBZ5, containing a wild-type copy of the A. parasiticuspyrG gene, was used to complement AF293.1, creating the reconstitutedstrain TJW55.2 (3). Extraction of fungal DNA, restriction enzymedigestion, gel electrophoresis, Southern and Northern blotting,hybridization, and probe preparation were performed using standardmethods (21, 31).

RNA blots were hybridized with a 3-kb PCR fragment of laeA usingprimers FUM1 (5'GCGCACTTCTTTGTTTCCCT 3') and FUM4 (5'CATGACGGTAACTAAGGATTTGG3'), a 1-kb alb1 PCR product using primers alb1f (5' AATCACTACTCCTTGCAGTCACGG3') and alb1r (5' GAGCCACCTTTAGCAGATCACGC 3'), a 1-kb rodA PCRproduct using primers rodAF (5'TGCATACCACATTCCACTCAGTG 3') androdAR (5' ACC TGT TCC TAC CCA GCT GCC 3'), and a 1-kb rodB PCRproduct using primers rodBF (5' AAC TAC AGC ACC GAA GAA GGTTG 3') and rodBR (5' TAA CGC TCC TCA AGT CCA ATA GG 3').

Phenotypic characterization and secondary metabolite analysis. Mycelia were cultured overnight in liquid GMM or RPMI 1640 (37°C),and the fungal mats were lyophilized and weighed. Colony growthradius, morphology, spore germination, and spore productionwere measured as previously described (15). Published procedureswere used to extract compounds from conidial spores of wild-typeA. fumigatus and the ${Delta}$ laeA mutant (26). Gliotoxin productionwas measured by modification of a previously published method(3). Secondary metabolites were measured from organic extractsof 7-day-old agar cultures (GMM) analyzed by liquid chromatography-photodiode array detection-high-resolution mass spectrometry (28)using the column and solvent system described by Bruhn et al.(6). Gliotoxin and other metabolites were identified by theirretention times, UV spectra, and positive electrospray spectra,which should match both the relative intensities of ions andtheir respective accurate masses (less than ±0.01 Daof deviation). Metabolites were measured in murine lung homogenateafter infection (methods detailed below).

Animal model of Aspergillus infection. The virulence of isogenic wild-type, pyrG, and ${Delta}$ laeA strainswas studied in a lung infection model, with approval of theUniversity of Wisconsin Animal Care Committee. Conidia wereharvested by flooding fungal colonies with 0.85% NaCl with Tween80, enumerated with a hemocytometer, and adjusted to a finalconcentration of 6.2 log₁₀ CFU/ml. Counts and the viabilityof the inocula were verified by duplicate serial plating onGMM plates. Six-week-old, outbred Swiss ICR mice (Harlan SpragueDawley), weighing 24 to 27 g, were immunosuppressed by intraperitonealinjection of cyclophosphamide (150 mg/kg of body weight) ondays –4, –1, and 3 and with a single dose of cortisoneacetate (200 mg/kg). The mice were anesthetized via halothaneinhalation in a bell jar on day 0. Sedated mice (10 mice/fungalstrain) were infected by nasal instillation of 50 µl ofthe inoculum (day 1) and monitored three times daily for 7 dayspostinfection. All surviving mice were sacrificed on day 7.The tissue fungal burden in whole-lung homogenate was quantifiedby serial dilution and enumeration of CFU (number of CFU/twolungs). The duration of survival in days after inoculation wasrecorded for each animal. Moribund animals were sacrificed,and cumulative survival was recorded. Survival and clearanceof residual fungal burden in tissue (number of CFU/two lungs)were used as the outcome variables to assess the relative virulencelevels of isogenic strains.

Human macrophage phagocytosis and killing of conidia. Peripheral blood mononuclear cells (PBMCs) from healthy adulthuman donors were isolated by centrifugation of heparinizedblood over a Ficoll-Hypaque density gradient at 800 x g for30 min. Isolated PBMCs were washed and resuspended to a concentrationof 2 x 10⁶ PBMCs/ml in complete medium. Monocytes were isolatedby adherence using Falcon Primaria-coated (24-well) tissue cultureplates (Becton Dickinson, Franklin Lakes, NJ). The cells thatadhered after 3 h at 37°C were separated and resuspendedin 1.5 ml of complete medium containing 10% human serum and40 ng/ml macrophage colony-stimulating factor. After 7 to 10days in culture, macrophages were documented to be >95% CD14⁺by flow cytometry. For isolation of murine thioglycolate-elicitedperitoneal macrophages, 1 ml of 4% Brewer-modified thioglycolatemedium (Becton Dickinson, Irvine, CA) was injected intraperitoneallyinto 12-week-old C57BL/6 mice. Five days later, peritoneal exudatecells, consisting of mostly macrophages (>90%), were collectedin 12 ml of phosphate-buffered saline (PBS). Macrophages werewashed with complete medium and resuspended to a concentrationof 5 x 10⁶ cells/ml. Cells were plated in Falcon Primaria-coated(24-well) tissue culture plates and incubated overnight at 37°Cwith 5% CO_2. Nonadherent cells were removed, and the purityof adherent macrophages was assessed by flow cytometry usingboth FITC-conjugated rat anti-mouse Mac-3 monoclonal antibodyand FITC-conjugated rat anti-mouse CD14 monoclonal antibody(BD Biosciences). For phagocytosis experiments, WT and ${Delta}$ laeAconidia were prelabeled with FITC by gentle shaking (30°C)in the dark for 1 h. Labeled conidia (conidia_FITC) were washedtwice with PBS, resuspended in RPMI 1640-HEPES, and used immediately.Macrophages were incubated with conidia_FITC at a ratio of 1:1,and plates were centrifuged briefly (500 x g) to ensure particleexposure. Macrophages were allowed to ingest conidia_FITC for5, 10, and 30 min at 37°C, and adherent cells were gentlywashed with PBS, counterstained with calcofluor white (25 µM),and reincubated for 15 min at 4°C in the dark. The cellswere then washed, fixed with 4% paraformaldehyde, and viewedwith a Deltavision wide-field fluorescence microscope (AppliedPrecision Inc., Issaquah, WA) that was fitted with an FITC andrhodamine filter set. The number of macrophages containing FITCgreen (internal) and calcofluor white blue (external adhered)organisms was quantified microscopically. The phagocytosis indexwas calculated as the percentage of phagocytic macrophages multipliedby the mean number of organisms internalized per macrophage.The percentage of phagocytosis was defined as the ratio of phagocytizingmacrophages that ingested one or more conidia. Macrophage viabilityafter conidial exposure was assessed by propidium iodide (PI;final concentration, 1 µg/ml) uptake after a 15-min incubationusing a fluorescence microscope. Experiments were performedwith triplicate wells and repeated at least three times. Macrophagekilling of conidia was measured by serial dilution as previouslydescribed (22). Monocyte-derived macrophages were allowed toingest conidia for 1 h, wells were washed with warm PBS, andcells were incubated for another 6 h before intracellular conidiawere harvested by exposure to isotonic saline with rapid freezing-thawing(–70°C to 37°C). Cellular lysis was confirmedby microscopic visualization, and the serial dilutions of supernatantwere plated on yeast extract agar containing Triton (10 g yeastextract, 20 g peptone, 20 g dextrose, 20 g agar, and 50 ml TritonX-100 per liter). Colonies were counted after 24 h of growthat 37°C.

PMN killing and fungal cytotoxicity. PMNs were purified from whole blood of healthy adult human donorsby Ficoll centrifugation, by following the leukocyte separationprotocol (Sigma). Briefly, red blood cells were lysed in hypotonicsolution, and PMNs were separated by centrifugation (800 x g,10 min) and resuspended in RPMI 1640-HEPES. Cellular viabilitywas assessed by trypan blue exclusion. PMN killing of Aspergillushyphae was measured by a modification of a previously describedXTT [2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide]assay (25). Briefly, conidial suspensions were prepared by floodingthe surface of the colonies with RPMI 1640-HEPES buffer andfiltered through gauze to separate the larger hyphal fragments.Early hyphal elements were generated from conidia (10⁶/ml) in96-well flat-bottom plates (Costar, Cambridge, MA) after incubationfor 8 h at 37°C. Under these conditions, more than 85% ofthe conidia germinated into germ tubes that were 4 to 5 timesthe length of the conidia. A calcein release assay was developedto measure the ability of live fungal hyphae to induce PMN damage.Calcein AM is an esterase dye that passively crosses the cellmembrane and is converted by intracellular esterases into aninsoluble fluorescent product that is retained by cells withintact membranes but released by damaged ones, producing anintense green signal (27). Wells containing hyphal elements(prepared as described above) were washed twice with prewarmed37°C medium, resuspended in fresh medium containing calceinAM (25 µM), and incubated at 37°C for 30 min. Onehundred microliters of the calcein-tagged PMNs were added tothe wells containing fungi (effector-to-target cell ratio, 1:1).Plates were incubated for 3 h at 37°C (5% CO₂) and centrifuged,and supernatants were transferred to 96-well FluoroNunc plates(Nalge Nunc, Denmark). Mean fluorescent units of calcein releasedinto the supernatant were measured using a fluorescence microplatereader (Victor² 1420-011 multilabel counter; Wallac, Inc., MD)(excitation filter, 485 ± 9 nm; band-pass filter, 530± 9 nm). Triplicate wells and live and dead (Triton X-treated)PMN controls were included for each experiment. Experimentswere repeated three times with different donors. PMN viabilitywas measured after exposure to concentrated culture supernatantby propidium iodide uptake. To prepare culture supernatant,40 ml GMM was inoculated with 10⁷ conidia/ml. After 3 days ofincubation at 37°C (with shaking at 300 rpm), culture mediumwas collected by filtration through sterile gauze. Liquid nitrogenwas added, and the product was vacuum dried for 36 h. The precipitatewas resuspended in 4 ml PBS and frozen at –70°C forlater use. PMNs (4 x 10⁵ cells), obtained as described above,were exposed to 200 µl concentrated supernatant, medium,or Triton X-100 for 5, 15, and 60 min. Cells were centrifuged(1,200 rpm, 10 min) and resuspended in PI (10 µg/ml) for15 min. The percentage of PI-positive (dead) cells was quantifiedby flow cytometry.

Conidial phenotypes determined by scanning electron microscopy and adherence assays. For scanning electron microscopy studies, conidia from 7-day-oldcultures were fixed in 2% glutaraldehyde for 60 min, 100 µlof the fixative solution was collected onto a 0.6-mm-pore-sizeNuclepore filter, and fixation was continued for an additional60 min. The filters were washed three times in double-distilledwater over a period of 10 min. Samples were postfixed with 1%osmium tetroxide for 50 min and washed three times. The filterswere dehydrated, mounted on aluminum stubs, and coated withgold-palladium alloy. An Amray 1820D scanning electron microscopewas operated at an accelerating potential of 25 kV to view theconidia. A microsphere adhesion assay was performed to assessrelative adhesive properties of conidia (9). Conidia (1 x 10⁸/µl)were resuspended in 0.1 mol/liter KNO₃ solution (pH 6.5). Latexparticles (0.8 µm; Sigma) were suspended with A. fumigatusconidia at a 20:1 ratio, and the number of microspheres adherentto each conidium was counted using light microscopy.

RESULTS

Top

Results

Phenotypic characterization. Despite the decreased production of mycelial metabolites, the ${Delta}$ laeA mutant demonstrated normal conidial pigmentation (3). Filamentousgrowth rates in liquid and colony cultures were similar between ${Delta}$ laeA mutant and WT isolates, although the A. fumigatus ${Delta}$ laeAmutant produced substantially more mass than the WT did underthe conditions tested (Table 2 and Fig. 1). The A. fumigatus ${Delta}$ laeA mutant was impaired in production of conidiophores andconidia in liquid shake culture, although the mutant's germinationof conidia was equivalent to that of the WT. Conidial productionlevels were similar on solid medium (3). The A. fumigatus ${Delta}$ laeAmutant-complemented strains presented the wild-type phenotypein all aspects of physiological growth (Fig. 1 and data notshown). Neither the A. nidulans WT nor the ${Delta}$ laeA mutant sporulatedin liquid shake culture (data not shown).

View this table:
[in this window]
[in a new window]
TABLE 2. Growth of the A. fumigatus ${Delta}$ laeA mutant

View larger version (107K):
[in this window]
[in a new window]
FIG. 1. Filamentous growth and conidiophore development of A. fumigatus WT, ${Delta}$ laeA mutant, and ${Delta}$ laeA mutant-complemented strains in liquid medium. Images obtained by microscopy of colonies developed after 48 h of incubation in shaking (300 rpm) culture at 25°C.

Animal model of Aspergillus infection and secondary metabolite analysis. The relative virulence levels of the WT and the A. fumigatus ${Delta}$ laeA mutant were evaluated in a murine pulmonary infection model(Fig. 2 and see Fig. S1 in the supplemental material). The nullmutant caused less infection than WT strains carrying laeA (WTand the pyrG strain) (Fig. 2) or the complemented strain (seeFig. S1 in the supplemental material). Fewer ${Delta}$ laeA fungal CFUwere recovered from the lungs of mice that were sacrificed 3days after infection (Fig. 2A and B). The isolated fungal coloniesfrom these lungs were confirmed to be the ${Delta}$ laeA mutant or WTby phenotype and Southern analysis (data not shown).

View larger version (32K):
[in this window]
[in a new window]
FIG. 2. Animal model of invasive aspergillosis. Outbred Swiss ICR mice immunosuppressed by intraperitoneal injection of cyclophosphamide (150 mg/kg) and cortisone acetate (200 mg/kg) were inoculated intranasally with WT, ${Delta}$ laeA, and pyrG conidia (50 µl of 5 x 10⁶ conidia/ml). (A) Survival of mice challenged with Aspergillus strains. The log rank test was used to perform pair-wise comparisons of survival levels among the strain groups. P value for comparison of the survival levels of WT- and ${Delta}$ laeA mutant-infected mice, 0.002. (B) Fungal burden, measured as number of CFU recovered from homogenized lungs. The Kruskal-Wallis one-way analysis of variance on ranks test was used to compare mean numbers of CFU among strain groups. P value for comparison of WT and pyrG strain CFU counts, >0.05. P value for comparison of ${Delta}$ laeA mutant and WT or pyrG strain CFU counts, <0.001. (C) Aspergillus metabolites present in lungs after infection. Chloroform extracts from mouse lungs were cleaned by reverse-phase solid-phase extraction and analyzed by liquid chromatography with diode array detection. The 200- to 600-nm chromatograms are inserted along with the UV spectrum of gliotoxin in the sample and the matching reference standard (upper dotted line) of gliotoxin. The deviation in retention time of gliotoxin (retention time, 6.65 min) was less than 0.01 min. The presence of gliotoxin was confirmed by liquid chromatography-phote diode array detection-high-resolution mass spectrometry (data not shown). mAU, milliabsorbance units.

The results of liquid chromatography-high-resolution mass spectrometryshowed that gliotoxin was present in lungs infected with WTand pyrG A. fumigatus isolates but not in ${Delta}$ laeA mutant-infectedor noninfected lungs (Fig. 2C). No other gliotoxin analoguesor secondary metabolites normally produced during in vitro cultureof A. fumigatus (e.g., helvolic acids, fumigaclavines A to C,tryptoquivalines A to H, fumiquinazolines A to E, pseurotinsA to F2, fumagillin, fumitremorgins A to C, verruculogen, orTR-2) could be detected in the WT- or ${Delta}$ laeA mutant-infected lungtissues (data not shown). ${Delta}$ laeA mutant-complemented strains wererestored in metabolite production (see Fig. S2 in the supplementalmaterial).

Macrophage phagocytosis. We reasoned that the decrease in pulmonary infection might bereflected by aberrancies in the fungus-macrophage interaction.Toward this end, the ability of macrophages to ingest ${Delta}$ laeA conidiawas measured. Human monocyte-derived macrophages were exposedto FITC-labeled conidia for different time periods, and fungalcells were counterstained with calcofluor white to distinguishbetween ingested and adherent cells. The number of macrophagescontaining FITC green (internal) and calcofluor white blue (externaladhered) organisms was quantified microscopically. Figure 3A and Bshow that macrophages internalized more ${Delta}$ laeA conidia thanWT conidia, with phagocytosis indices separating within 10 minof exposure. Similar results were observed with murine-derivedmacrophages (data not shown). Altered phagocytosis was not associatedwith differences in viability of exposed macrophages, as measuredby propidium iodide uptake (data not shown). No differencesin intracellular killing of ${Delta}$ laeA and WT conidia were measuredby either serial dilution plating or intracellular germination(data not shown); however, the large differences in the numbersof organisms ingested per macrophage likely limited our abilityto detect small differences in intracellular killing.

View larger version (17K):
[in this window]
[in a new window]
FIG. 3. Conidial and hyphal interactions with human phagocytic cells. (A and B) Human CD14⁺ monocyte-derived macrophages ingested higher numbers of ${Delta}$ laeA conidia than of WT conidia (internal conidia are stained green). The phagocytosis index was calculated as the percentage of phagocytic macrophages multiplied by the mean number of organisms internalized per cell. Experiments were performed with triplicate wells and repeated at least three times.

Conidial phenotype. To define the nature of the differences in macrophage phagocytosis,we more closely examined spore metabolites and architecture.Conidia harvested from the ${Delta}$ laeA mutant produced diminished theamounts of at least one metabolite in vitro (Fig. 4A). Surfacefeatures of conidia harvested from the ${Delta}$ laeA mutant demonstratedoverall losses of prominent protrusions, consistent with defectiverodlet synthesis (Fig. 4B). Accordingly, ${Delta}$ laeA conidia were lessadherent to latex microspheres than the WT conidia were (Fig.4C), and Northern profiling revealed delayed expression of genesencoding rodlets (rodA and rodB) and conidial pigments (alb1),which was correlated with sporulation patterns (Fig. 4D). Thesedata indicate that LaeA regulates the expression of genes involvedin conidial biosynthesis.

View larger version (65K):
[in this window]
[in a new window]
FIG. 4. Conidial properties. (A) Spore metabolites analyzed by thin-layer chromatography. Spore diffusate was obtained by incubation of 10⁸ cells/ml in Hanks balanced salt solution for 1 h; products were filter sterilized, chloroform extracted, and dried under vacuum suction overnight. Products were redissolved with 100 µl chloroform for thin-layer chromatography as previously described (26). Metabolite produced by the WT and not the ${Delta}$ laeA mutant is indicated with an arrow. (B) ${Delta}$ laeA conidia (right panel) have decreased rodlet protrusions compared to WT conidia. (C) ${Delta}$ laeA conidia are less adhesive, as measured by adhesion of latex microspheres (9). Conidia (1 x 10⁸/µl) were suspended in 0.1 mol/liter KNO₃ solution (pH 6.5) and incubated at a 20:1 ratio with latex particles (0.6 µm); the percentage of conidia with adhered particles was calculated from triplicate measurements in two experiments. The two-tailed P value from Student's t test comparing results for the ${Delta}$ laeA mutant with those of the WT and the pyrG strain was <0.0001. The P value comparing results for the WT and the pyrG strain was >0.05. (D) Northern blots demonstrating laeA, alb1, rodA, and rodB gene expression in A. fumigatus ${Delta}$ laeA mutant and WT strains after growth in GMM from 24 to 72 h. Ethidium bromide-stained rRNA is indicated for loading controls.

PMN killing and fungal cytotoxicity. We next measured the ability of PMNs to kill A. fumigatus hyphaeas well as the ability of hyphae to kill neutrophils. A. fumigatusinvasive growth is via hyphal proliferation, where neutrophilsplay a role in host defense. We found that human PMNs were equallyeffective in killing WT and ${Delta}$ laeA hyphae 1 h after exposure,as measured by conversion of the viability dye 2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide(data not shown). Speculating that differences in PMN viabilitymay become apparent due to alterations in secondary metaboliteelaboration during hyphal growth of the ${Delta}$ laeA mutant, we measuredPMN death 3 h after exposure to live hyphae by using an assaythat measures the release of intracellular calcein AM (27). ${Delta}$ laeA hyphae induced less PMN cellular death than did hyphaeof the WT or the pyrG strain (Fig. 5A). During this late timeperiod, it is not possible to accurately measure hyphal viabilityor PMN killing of fungi due to fungal overgrowth. Differencesin PMN death are associated at least in part with Aspergillusproducts secreted during later stages of growth, as the samedifferences in cellular viability were observed after PMN wereexposed to concentrated WT and ${Delta}$ laeA mutant culture supernatants(Fig. 5B). Collectively, these data suggest that multiple phenotypesare associated with laeA deletion; decreased PMN death is consistentwith altered secondary metabolite production by hyphae.

View larger version (14K):
[in this window]
[in a new window]
FIG. 5. PMN killing and fungal cytotoxicity. (A) Damage to PMN (calcein release) after 3 h of incubation with live WT and ${Delta}$ laeA mutant germ tubes was measured by the release of calcein AM (25). Shown are mean fluorescent units (MFU) of triplicate measurements for six healthy donors (±standard deviations), representative of results obtained for nine different experiments. The two-tailed P value from Student's t test comparing the results for the WT and the pyrG strain was >0.05. The P value for comparisons of the results for the ${Delta}$ laeA mutant with the WT and the pyrG strain was <0.001. (B) PMN death after exposure to concentrated culture supernatant from the WT and the ${Delta}$ laeA mutant. The percentage of total cells that were PI positive is shown. Controls (not shown) included Triton X-treated PMNs and untreated PMNs, which were >90% and <10% dead, respectively, with the assay. Results are means of triplicate measurements ± standard deviations from two experiments with different donors.

DISCUSSION

Top

Discussion

In its environmental niche, Aspergillus species propagate ina multicellular mycelium; at a critical cell density, the organismgenerates aerial stalks with terminal conidia produced on conidiophores.Conidia are readily aerosolized and are the infective form ofthe organism; in the lungs, conidia that escape killing by residentphagocytes germinate into hyphae, which can cause local pneumoniaand disseminate hematogenously (17). What distinguishes Aspergillusfumigatus from the remaining hundreds of thousands of filamentousmolds is not clear, as few single-gene "virulence factors" havebeen identified (4). Microbial disease attributes include acombination of factors secreted from growing mycelia and terminalhyphal cells and cell wall structural components, includinghydrophobins and pigments that confer resistance to phagocytickilling (11-13, 17, 34-37). The impaired virulence of the ${Delta}$ laeAstrain correlates with the loss or alteration of several ofthese factors of pathogenicity.

Our findings are the first genetic studies to draw potentialsignificance to the role of A. fumigatus secondary metabolitesin the establishment of intrapulmonary infection (Fig. 2; seeFig. S1 in the supplemental material). We hypothesized thatsecreted cytotoxic molecules may impact the relative abilityof host phagocytic cells to ingest and kill fungal conidialand hyphal elements ex vivo. The increased macrophage phagocytosisof ${Delta}$ laeA conidia was correlated with the loss of at least onesecreted factor (Fig. 4A). Although not well defined, factorsdiffused from spores have been reported to impact macrophagehandling of A. fumigatus conidia (26), including inhibitionof phagocytosis (2). Characterization of the metabolites missingfrom ${Delta}$ laeA conidia may provide insight into the factors governingphagocytosis.

Another factor possibly contributing to the increased phagocytosisof ${Delta}$ laeA conidia was the altered architecture of the spore walls(Fig. 4B). The outermost cell wall layer of Aspergillus conidiais decorated with interwoven proteinaceous microfibrils calledrodlets. These are hydrophobins that confer physiochemical propertiesto mediate conidial dispersal and, possibly, cellular interactions(9, 33). Conidial pigments, encoded by developmentally regulatedclustered genes, confer resistance to macrophage phagocytosisand intracellular killing (14, 34, 36, 37). Our current dataindicate that LaeA regulates the expression of genes involvedin both processes. Both alb1 and rodA null mutants are moresensitive to macrophage killing (19, 29). However, it is unknownwhether misscheduled expression of any of these genes couldalter conidium-macrophage interactions, a topic of further explorationin our lab. Interestingly, our data may suggest that adherenceis not an important factor in phagocytosis, as the ${Delta}$ laeA conidia,although more easily ingested by macrophages, adhered less wellto microspheres (Fig. 4C).

Whereas macrophage-mounted defense is most important for clearingconidial burden, invasive growth of A. fumigatus hyphae is counteredthrough host production of neutrophils. Here our experimentationstrongly implicated the role of secreted metabolites in thekilling of these host cells (Fig. 5B) with emphasis on the potentialrole of gliotoxin (Fig. 2C) in this process. Access to the A.fumigatus genome (www.tigr.org) indicates the potential forthe presence of up to 40 novel metabolites from polyketide andnonribosomal peptides alone, and the characterization of a putativegliotoxin gene cluster may shortly answer the question of howmuch this hyphal toxin contributes to the phenotype describedherein.

Genes encoding conidial components such as pigments and hydrophobinsare positively regulated during sporulation to allow for theproduction of conidia that are resistant to environmental stresses(33, 36); secondary metabolites are considered part of the chemicalarsenal required for niche specialization (7). The data presentedwithin show that LaeA acts as a regulator of gene expressionduring vegetative growth and sporulation in A. fumigatus; disruptionof this positive regulator yields a mutant with reduced abilityto establish pulmonary disease under the conditions described.This relatively avirulent phenotype is associated with alteredexpression of multiple microbial determinants critical for conferringboth conidial and hyphal cellular resistance to host phagocyticattack. Given the opportunistic nature of these organisms, itis most likely that virulence is dependent on coordinate expressionof multiple gene products that mediate host-fungus interactions.As A. fumigatus full-genome arrays will soon be available, itmay be possible to identify additional conidial and hyphal genesinvolved in conferring microbial resistance to host defensesand the ability to cause invasive and allergic disease in ${Delta}$ laeAmutant-profiling experiments. Figure 6 depicts our current understandingof LaeA regulation of A. fumigatus virulence attributes.

View larger version (24K):
[in this window]
[in a new window]
FIG. 6. Proposed model of LaeA regulation of A. fumigatus phenotypes that confer virulence. LaeA regulation by the protein kinase A pathway was previously reported (13). Virulence factors purported but not yet identified by genetic analyses (spore diffusible metabolite) are shown with dashed lines.

The implications of these findings extend well beyond the fieldof medical mycology. Data suggest that LaeA serves as a regulatorfor pathogenicity genes during fungal sporulation and hyphalelongation, critical morphological events shared by, but specificto, filamentous plant and animal pathogens. In A. nidulans,LaeA was shown to be a regulator for secondary metabolite production,including products with pharmaceutical and toxic properties(3). Genes homologous to laeA are found in other pathogens thatexhibit a filamentous cellular state, including Coccidioidesimmitis (GenBank), Fusarium sporotrichioides (http://www.genome.ou.edu/fsporo.html),and Magnaporthe grisea (http://www-genome.wi.mit.edu/annotation/fungi/magnaporthe).These findings indicate that LaeA is a critical developmentalregulator of genes involved in fungal environmental propagationand disease in plants and animals.

ACKNOWLEDGMENTS

This research was funded by NIH R01 51468 to K.A.M. and NSFMCB-0196233 and NSF MCB-0236393 to N.P.K.

FOOTNOTES

* Corresponding author. Mailing address: Univ. of Wisconsin—Madison, Department of Plant Pathology, 882 Russell Labs, 1630 Linden Drive, Madison, WI 53706. Phone: (608) 265-9795. Fax: (608) 263-2626. E-mail: npk{at}plantpath.wisc.edu.

Supplemental material for this article may be found at http://ec.asm.org/.

These authors contributed equally to this work.

REFERENCES

Top