Morphological Transitions Governed by Density Dependence and ...

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Sept. 2008, p. 5674–5685 0099-2240/08/$08.00⫹0 doi:10.1128/AEM.00565-08 Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Vol. 74, No. 18

Morphological Transitions Governed by Density Dependence and Lipoxygenase Activity in Aspergillus flavus䌤† S. Horowitz Brown,1 R. Zarnowski,2 W. C. Sharpee,1 and N. P. Keller1,2* Department of Plant Pathology1 and Department of Medical Microbiology and Immunology,2 University of Wisconsin—Madison, 1550 Linden Dr., Madison, Wisconsin 53706-1598 Received 9 March 2008/Accepted 10 July 2008

Aspergillus flavus differentiates to produce asexual dispersing spores (conidia) or overwintering survival structures called sclerotia. Results described here show that these two processes are oppositely regulated by density-dependent mechanisms and that increasing the cell density (from 101 to 107 cells/plate) results in the lowest numbers of sclerotial and the highest numbers of conidial. Extract from spent medium of low-celldensity cultures induced a high-sclerotium-number phenotype, whereas high-cell-density extract increased conidiation. Density-dependent development is also modified by changes in lipid availability. Exogenous linoleic acid increased sclerotial production at intermediate cell densities (104 and 105 cells/plate), whereas oleic and linolenic acids inhibited sclerotium formation. Deletion of Aflox encoding a lipoxygenase (LOX) greatly diminished density-dependent development of both sclerotia and conidia, resulting in an overall increase in the number of sclerotia and a decrease in the number of conidia at high cell densities (>105 cells/plate). Aflox mutants showed decreased linoleic acid LOX activity. Taken together, these results suggest that there is a quorum-sensing mechanism in which a factor(s) produced in dense cultures, perhaps a LOX-derived metabolite, activates conidium formation, while a factor(s) produced in low-density cultures stimulates sclerotium formation. spores) to conidial production is balanced by oxylipin availability. Oxylipins represent a vast and diverse family of secondary metabolites that originate from the oxidation or further conversion of unsaturated fatty acids. In A. nidulans endogenous oleic acid-, linoleic acid-, and likely linolenic acid-derived oxylipins, collectively called “psi factors” (precocious sexual inducer), influence the development of cleistothecia and conidia (6, 7, 8, 9, 51). The proportion of these different oxylipins has been postulated to regulate the ratio of asexual development to sexual development in A. nidulans. Such a “balancing” role for oxylipins was genetically supported by the results of deletion of the dioxygenases responsible for psi factor production. Loss of ppoB (psi factor-producing oxygenase B) yielded an increased-conidium, decreased-cleistothecium phenotype (52), whereas loss of both ppoA and ppoC resulted in the opposite phenotype (51). Additionally, exogenous applications of oxylipins resulted in phenotypes that mimicked the ppo mutant phenotypes in A. nidulans and also altered sclerotial and conidial production in A. flavus (7). Results of these studies suggested a possible quorum-driven, cell density-dependent phenomenon in morphological transitions in Aspergillus species (14, 50). Cell density-dependent regulatory networks in microorganisms generally control processes that involve cell-cell interactions, such as group motility and the formation of multicellular structures leading to differentiation processes (1, 25). Most cell density studies have centered on quorum sensing in bacteria (25). The swarm motility of microorganisms such as Vibrio parahaemolyticus and Proteus mirabilis is a multicellular behavior dependent on cell density (1). Likewise, the marine luminescent bacterium Vibrio fischeri uses the LuxR and LuxI proteins for autoinduction of luminescence (15). However, cell density-dependent phenomena are not limited to prokaryotes. Dictyostelium discoideum, a simple eukaryote, exhibits quorum-

Aspergillus flavus is a cosmopolitan, soilborne, filamentous fungus that frequently infects oil-rich seeds of several crop species. Concerns about the association of this organism with plant products, particularly corn, cotton, peanuts, and tree nuts, center on its ability to produce aflatoxins, which are carcinogenic polyketide secondary metabolites (3, 13). In addition, A. flavus is second only to Aspergillus fumigatus as a cause of human invasive aspergillosis (22). Asexual spores (conidia) are the primary disseminating mechanism of A. flavus and serve as the major inoculum source. A. flavus also differentiates to produce sclerotia, which are compact masses of mycelia that overwinter and germinate to produce either additional hyphae or conidia (13). Sclerotia are hypothesized to be degenerate sexual structures and may represent a vestige of cleistothecium production (16). Sclerotia allow A. flavus to survive under harsh conditions and to outcompete other organisms for substrates in the soil or in plants (13). Infected plant tissue, such as corn kernels, cobs, and leaf tissue, can remain in the soil and support the fungus until the following season, when newly exposed mycelium or sclerotia can give rise to conidial structures, thus producing the primary inoculum for the next infection cycle (13). In the genetic model organism Aspergillus nidulans, which does not produce sclerotia but does produce cleistothecia, the ratio of cleistothecial production (and production of asco-

* Corresponding author. Mailing address: 3476 Microbial Science Building, Department of Medical Microbiology and Immunology and Department of Plant Pathology, UW—Madison, 1550 Linden Dr., Madison, WI 53706. Phone: (608) 262-9795. Fax: (608) 262-8418. E-mail: [email protected]. † Supplemental material for this article may be found at http://aem .asm.org/. 䌤 Published ahead of print on 25 July 2008. 5674

VOL. 74, 2008

QUORUM SENSING IN A. FLAVUS

TABLE 1. A. flavus strains used in this study Strain(s)

a

NRRL 3357 NRRL 3357.5 TJSPara17 TSHB2.32, TSHB2.39, and TSHB2.40 TSHB3.5C, TSHB3.1C, and TSHB3.2(3)C a

Genotype

Reference

Wild type pyrG⫺ pyrG; pyrG⫺ ⌬Aflox::pyrG; pyrG⫺

41 41 This study This study

Aflox::phleomycin; ⌬Aflox::pyrG; This study pyrG⫺

Strains whose designations begin with T are original transformants.

sensing behavior. When an adequate bacterial food source is present, Dictyostelium cells live as unicellular amoebae and divide by fission. However, when food is scarce, Dictyostelium cells enter a developmental cycle that begins with the aggregation of cells that behave as a multicellular organism (31). Much research has also focused on quorum sensing in the fungus Candida albicans. In this organism at least two molecules, including the oxygenated lipid farnesol, regulate quorum sensing, yeast-to-hypha transitions, and biofilm formation (11, 38, 41). Density-dependent developmental transitions have been found in other fungi; for example, Ceratocystis ulmi grows as a yeast at high densities but as filaments at low densities (26). The switch can be attenuated by chemical inhibition of lipoxygenase (LOX) activity (28), indicating that oxylipins have a possible role in quorum sensing in C. ulmi. Here we examined cell density transitions in A. flavus, focusing on conidial and sclerotial development of the fungus grown at different cell densities. We found that both of these processes are regulated by density-dependent mechanisms and that increasing the cell density (from 101 to 107 cells/plate) resulted in the lowest numbers of sclerotia, yet the highest numbers of conidia. Extracts from spent medium of cultures with low cell densities stimulated sclerotial formation, and extracts from spent medium of cultures with high cell densities stimulated conidial formation, suggesting that there is a quorum-sensing mechanism in A. flavus. Furthermore, these density-dependent phenomena are influenced by lipid modifiers as both exogenous fatty acid application and loss of a putative oxylipin-generating LOX, encoded by Aflox, affect the conidiumto-sclerotium switch. MATERIALS AND METHODS Fungal strains and growth conditions. The A. flavus strains used and generated in this study are shown in Table 1. All strains used for physiological studies were prototrophic, and the strains were grown at 29°C and maintained on glucose minimal medium (GMM) (47) unless otherwise indicated. Cultures were grown either in the dark inside cardboard boxes that provided continuous darkness or in the presence of continuous white light by using an incubator equipped with a General Electric 15-W broad-spectrum fluorescent light bulb (F15T12CW) positioned 50 cm from the agar surface. Assay for conidial and sclerotial density dependence. For analysis of sclerotial density dependence, 10 ml of GMM with 1.6% agar and 2% sorbitol was overlaid with 3 ml of GMM with 0.7% agar and 2% sorbitol containing 101 to 107 conidia of an appropriate strain in a petri plate (60 by 15 mm). The plates were incubated at 29°C under continuous dark conditions for 7 days. To accurately visualize the sclerotia, plates were sprayed thoroughly with 70% ethanol in water to remove conidia and aerial mycelia. Either the exposed sclerotia then were counted with a dissecting scope at a magnification of ⫻4 or the sclerotia were collected and lyophilized and sclerotium production was determined gravimetrically (mg [dry weight] per plate). Conidia were counted on day 3 or 7. To accurately count conidia, three 1-cm plugs were homogenized in 3 ml of water containing 0.01%

5675

Tween 80, diluted 1:10, and counted with a hemocytometer at a magnification of ⫻400. For each conidial concentration, conidial and sclerotial counts were determined by using four replicates, and tests were repeated three times in order to examine reproducibility. Spore suspensions were freshly prepared for each strain and replication. Spent medium bioassay. For preparation of cell extracts from spent medium of A. flavus wild-type strain NRRL 3357 cultures, 30 petri plates (100 by 15 mm), each containing 25 ml of GMM supplemented with 1.6% agar plus 2% sorbitol, were overlaid with 10 ml of GMM supplemented with 0.7% agar plus 2% sorbitol containing 103 or 107 conidia/plate. Thirty uninoculated agar plates were used as the control. All 90 plates were incubated at 29°C under continuous dark conditions for 5 days, after which the cultures were extracted with ethyl acetate as follows. For convenience, the contents of 10 plates were extracted at a time. The agar and fungus in 10 plates were removed using a spatula, placed in a large blender with 250 ml of sterile distilled H2O, and blended for 10 s. The macerate was transferred into a 4-liter beaker, and the procedure described above was repeated for the remaining 20 plates, 10 at a time, for each treatment. The total macerate was placed into the fume hood, and 2 volumes of ethyl acetate was added. The macerate with the ethyl acetate was stirred with a magnetic stirrer for 15 min, allowing the layers to separate. The organic layer was then removed using a 100-ml glass pipette and placed into a 1-liter evaporating flask (Rotavapor-RE; Brinkmann, Switzerland). The dry extract was resuspended in a few milliliters of ethyl acetate and transferred into a glass vial. The vial was covered with foil (to protect the contents from light), and the extract was again evaporated and resuspended in 2.5 ml ethanol, topped with nitrogen gas, and kept at ⫺80°C until it was used. Extracts of spent medium were added to new cultures by spreading 100-␮l portions of 1:5, 1:10, or 1:100 dilutions (in ethanol) from the 2.5-ml ethanol extracts from the plates containing 103 or 107 conidia/plate and medium extracts obtained as described above on petri plates (60 by 15 mm) containing 10 ml of GMM supplemented with 1.6% agar plus 2% sorbitol. A water control was also included. The plates were allowed to dry for 15 min and then overlaid with 3 ml of GMM supplemented with 0.7% agar plus 2% sorbitol containing 101 to 107 conidia. This yielded a total of 112 petri plates for the entire experiment. Cultures were allowed to grow at 29°C under continuous dark conditions for 7 days. To accurately visualize the sclerotia, plates were sprayed thoroughly with 70% ethanol in water to remove conidia and aerial mycelia. The exposed sclerotia then were collected and lyophilized, and sclerotium production was determined gravimetrically (mg [dry weight] per plate). Numbers of conidia were determined only for the plates containing 106 and 107 conidia/plate treated with the low-density (103 conidia/plate), high-density (107 conidia/plate), and medium extracts at 3 days. Conidia were counted by using the method described above for the assay for conidial and sclerotial density dependence using four replicates for each combination. The entire experiment was repeated twice in order to examine reproducibility. Extracts from low- and high-density plates and medium extract were freshly prepared for each experiment. Fatty acid assay. Linoleic acid, oleic acid (Cayman Chemical, Ann Arbor, MI), arachidonic acid (TCI America, Portland, OR), and linolenic acid (MP Biomedicals, Solon, OH) were dissolved in ethanol at concentrations of 0.1 and 1 mg/60 ␮l and then applied to and dried on 12.5-mm-diameter filter paper disks. A filter paper disk treated with 60 ␮l of ethanol was used as the solvent control. After drying, the disks were laid on agar surfaces after plates were inoculated with 101 to 107 conidia/plate. Cultures were incubated in the dark inside cardboard boxes (resulting in continuous darkness) for 7 days. To accurately visualize the sclerotia, plates were sprayed thoroughly with 70% ethanol in water to remove conidia and aerial mycelia. Sclerotia were collected and freeze-dried, and sclerotium production was determined gravimetrically (mg [dry weight] per plate). All of the fatty acid tests with wild-type strain NRRL 3357 were performed by using three replicates, the entire experiment was repeated twice, and similar results were obtained. Nucleic acid analysis. Extraction of DNA from fungi, restriction enzyme digestion, gel electrophoresis, blotting, hybridization, and probe preparation were performed by using standard methods (46). To examine Aflox expression, cultures of wild-type strain NRRL 3357 and complemented strains TSHB3.1C and TSHB3.5C were cultured in shaken liquid GMM containing 2% yeast extract for 48 h at 29°C and 250 rpm. Equal amounts of mycelium were then removed, placed in liquid GMM containing 2% sorbitol, and grown for an additional 2 h with shaking 250 rpm at 29°C; this was followed by harvesting. Three separate repetitions of this experiment yielded similar results. Aflox expression was also assessed by growing strain NRRL 3357, ⌬lox strain TSHB2.39, and strain TSHB3.5C in liquid GMM containing 2% sorbitol for 24 h at 29°C and 220 rpm and then transferring the mycelia to filter paper on solid GMM containing 2%

5676

HOROWITZ BROWN ET AL.

sorbitol. Total RNA was extracted at the time of the shift (zero time) and 6, 12, 24, and 48 h after the shift to solid culture. Two separate repetitions of this experiment yielded similar results. Total RNAs were extracted from lyophilized mycelia using Trizol reagent (Invitrogen, Carlsbad, CA). RNA (10 to 30 ␮g) was separated on a 1.3% formaldehyde gel. RNA was transferred to a Hybond membrane (Amersham Bioscience, Piscataway, NJ) by capillary action. RNA transcript analysis was performed using random primer radiolabeled DNA fragments (46) derived from genomic PCR products. Aflox expression studies were performed with an Afloxspecific DNA probe (a 1.07-kb PCR fragment) which was generated from genomic DNA using the following primers: loxF (5⬘-AGGCCCGTATGAGCA GTTGAGT-3⬘) and loxR (5⬘-GTTACCTTTACGGCTCCCCTCT-3⬘). An actinspecific DNA probe (a 0.5-kb PCR fragment) was generated from genomic DNA using the following primers: ActinF (5⬘-ACAGTCCAAGCGTGGTATCC-3⬘) and ActinR (5⬘-GAAGCGGTCTGAATCTCCTG-3⬘). Signals were detected with a Phosphorimager-SI (Molecular Dynamics). Nucleotide sequences were analyzed and compared using Sequencher (Gene Codes) and ClustalW (www .ebi.ac.uk/clustalw/) programs (12). Deletion and complementation of the A. flavus lox gene. The Aflox gene was identified by a TBLASTX search based on a putative A. fumigatus arachidonic 5-lipoxygenase (accession no. XP_746463), which was used as a query sequence. Only one lox sequence was found, as confirmed by Southern analysis. The Aflox deletion construct pSHB2, including the Aspergillus parasiticus pyrG marker gene and Aflox flanking sequences, was constructed using the following method. First, the modified primers 5⬘ flank loxF SalI (5⬘-CCAGGTGTCGACAAAAAGGA GAAAGGAAGCAAA-3⬘) and 5⬘ flank loxR EcoRI (5⬘-TCCTTCGAGTACA CGTAGGGAATTCAGCCCG-3⬘) were used to PCR amplify a 1.07-kb flanking region at the 5⬘ untranslated region of the Aflox open reading frame using A. flavus genomic DNA as the template. The resulting amplified SalI-EcoRI PCR fragment was subcloned into pJW24 harboring the A. parasiticus pyrG cassette (5), yielding the vector pSHB1. Next, the modified primers 3⬘ Flank loxF XbaI (5⬘-TTATTGTCTAGAGATGTCTTCGAGATTTGAACC-3⬘) and 3⬘ Flank loxR SacI (5⬘-ATATTGTCATGTTTGGGACGGAGCTCTAATTG-3⬘) were used to PCR amplify a 1.3-kb flanking region at the 3⬘ end of the predicted Aflox open reading frame. The amplified XbaI-SacI 3⬘ flanking region was further ligated into XbaI-SacI-digested plasmid pSHB1, generating plasmid pSHB2. The final deletion vector, pSHB2, was used as a template to PCR amplify a fragment containing the Aflox 5⬘ flanking region, a pyrG selectable marker, and the Aflox 3⬘ flanking region with the following primers: Aflox nested F (5⬘-CAAGAGCA GTAGCAGCAGAAGG-3⬘) and Aflox nested R (5⬘-GGGCCCACTCACAAC GTATCAT-3⬘). The PCR product was used to transform A. flavus NRRL 3357.5 (pyrG⫺) to create Aflox deletion strains. Fungal transformation was performed essentially as described by Miller et al. (36). The Aflox deletion strains (⌬Aflox) were obtained by a double-crossover event, exchanging the pyrG selectable marker gene for the Aflox coding region (see Fig. 4A). Southern analysis was used to confirm that there was a single gene replacement event in TSHB2.39, TSHB2.40, and TSHB2.32 (Fig. 4B) with an Aflox full-length probe which was obtained by performing PCR with primers Aflox.comp F (5⬘-GCTGATATTCC GTCCAGTTCG-3⬘) and Aflox.comp R (5⬘-CTGGATTTGTCATCGTGCAG3⬘). Signals were detected with a Phosphorimager-SI (Molecular Dynamics). Complementation of the ⌬lox transformant TSHB2.39 was achieved using the vector pSHB3. Plasmid pSHB3 was created by inserting a 3.09-kb fragment containing the predicted promoter, coding sequence, and termination cassette of Aflox into pBC-Phleo. Plasmid pBC-Phleo carries the phleomycin resistance cassette in which the ble gene is under control of the A. nidulans gpdA promoter and the Saccharomyces cerevisiae CYC1 terminator (21). Phylogenetic analysis. Conservation of LOX proteins was searched using BLASTP. Hits for putative fungal, plant, and mammalian LOXs were aligned, and a phylogenetic tree was created using the ClustalW program. Pairwise scores for the amino acid sequences were calculated by dividing the number of identities in the best alignment by the number of residues compared (gap positions were excluded). The phylogenetic tree was calculated based on the multiple alignment, and the distances between the amino acid sequences in the alignment were then used by the TreeView software program (12) to construct the tree shown in Fig. S1 in the supplemental material. The GenBank accession numbers of the sequences of other fungal, plant, and mammalian LOXs which were used to generate the tree are as follows: A. fumigatus, XP_746463 and XP_746844; Neosartorya fischeri, XP_001262545; Gaeumannomyces graminis, AAK81883; Magnaporthe grisea, XP_362938; Botryotinia fuckeliana, XP_001550612; Gibberella moniliformis, AAW21637; Neurospora crassa, CAD37061; Chaetomium glabosum, XP_001225066; Homo sapiens 12-LOX, NP_001130; soybean, 1Y4K_A; Zea mays, NP_001105975; and Oryza sativa, NP_001055143.

APPL. ENVIRON. MICROBIOL. Lipid extraction and fatty acid analysis. Wild-type strain NRRL 3357, ⌬lox strain TSHB2.39, and complemented strain TSHB3.5C were grown in 10 ml of liquid GMM containing 2% sorbitol at a density of 105 conidia/plate under stationary conditions at 29°C in the dark. Mycelial mats collected after 120 h of growth were frozen in liquid nitrogen and ground with a mortar and pestle. Ground mycelial samples were then transferred to beakers and soaked in 30 ml of a chloroform-methanol mixture (1:1, vol/vol). Lipids were extracted for 48 h at ⫺20°C. Next, the extraction mixtures were filtered through filter paper with low porosity (Fisher Scientific) and transferred to 50-ml centrifuge tubes (Pyrex). Ten microliters of 0.97% KCl was then added to each tube, thoroughly mixed, and left to stand for 1 h. To separate the organic and aqueous phases, the tubes were centrifuged (300 ⫻ g, 5 min), and the top aqueous layer and the interphase were removed with a Pasteur pipette. The organic layer containing extracted lipids was washed with 2.0 ml of 0.97% KCl, and the top separated phase was again removed. The remaining chloroform extracts were dried with a stream of nitrogen, and the solid lipid residues were resuspended in 1 ml of chloroform and transferred into dark-glass vials. Fatty acids were converted into corresponding methyl ester derivatives in 2% sulfuric acid in methanol (Sigma). Prepared fatty acid methyl esters were extracted with n-hexane. All solvents contained 5 mg liter⫺1 butylated hydroxytoluene as an antioxidant. Fatty acid methyl esters were identified by gas chromatography using a Hewlett-Packard 5890 equipped with a capillary column coated with DB-225 (length, 30 m; 0.25 mm; internal diameter, 0.25 ␮m; Agilent Technologies, Inc., Wilmington, DE). The column temperature was kept at 70°C for 1 min and increased to 180°C at a rate of 20°C min⫺1 and then to 220°C at a rate of 3°C min⫺1. Then the temperature was kept at 220°C for 15 min. The injector and detector temperatures were set at 250°C, and the injection port temperature was set at 300°C. Peaks were identified by comparing retention times with the retention times of a set of authentic fatty acid standards provided by Supelco. The abundance of fatty acids was calculated from relative peak areas. Cell protein extraction and determination of LOX activity. Wild-type strain NRRL 3357, ⌬lox strain TSHB2.39, and complemented strains TSHB3.5C, TSHB3.1C, and TSHB3.2(3)C were cultured in shaken liquid GMM containing 2% yeast extract for 48 h at 29°C and 250 rpm, and equal amounts of mycelium were then transferred into liquid GMM containing 2% sorbitol, allowed to grow for an additional 2 h with shaking at 250 rpm at 29°C, and harvested. Mycelial balls were immediately collected and transferred into 2-ml screw-top plastic tubes containing chloroform-washed glass beads (diameter, ⱕ106 ␮m; Sigma, St. Louis, MO) and 1 ml of phosphate-saline buffer (pH 7.2). The mycelia were homogenized mechanically using a Mini Beadbeater-8 (Biospec Products Inc., Bartlesville, OK) for 1.5 min, followed by 3 min of cooling on ice. After this, the slurry was centrifuged (10,000 ⫻ g, 5 min, 4°C), and cell homogenates were transferred into fresh Eppendorf tubes. Protein contents of cell extracts were determined using a bicinchoninic acid protein assay kit (Pierce Biotechnology, Rockford, IL) with bovine serum albumin as a standard. LOX was assayed spectrophotometrically using a modified method of Axelrod et al. (2). The 10 mM substrate stock mixture contained 157.2 ␮l of linoleic acid and 157.2 ␮l of Tween 20 dissolved in 10 ml of water. The solution was thoroughly mixed, 1 ml of 1 M NaOH was added, and the volume was subsequently adjusted to 50 ml with water. Prior to the assay, the substrate stock mixture was diluted with phosphate-saline buffer (1:4) to obtain a final linoleic acid concentration of 2.5 mM. The solution was then bubbled with a 95% oxygen-5% CO2 mixture for 20 min and allowed to equilibrate for 15 min before use. For the LOX assay, 100 ␮l of crude cell extract was added to 500 ␮l of the working substrate solution, mixed well, and incubated in the dark for 20 min at room temperature. After this, the reaction mixture was extracted with 700 ␮l of a chloroform-methanol mixture (1:1, vol/vol) and vortexed, and the top aqueous layer and the interphase were removed after centrifugation (300 ⫻ g, 5 min). The organic layer was washed with 200 ␮l of 0.97% KCl, and the top separated phase was again removed. The remaining chloroform extracts were dried with a stream of nitrogen, and the solid lipid residues were resuspended in 500 ␮l of chloroform and transferred to a 1-cm-path-length quartz cuvette. The formation of conjugated diene was measured at 234 nm with a Beckman DU530 Life Science UV/VIS spectrophotometer. One unit of specific enzymatic activity was defined a change in absorbance at 234 nm of 1 milli-absorbance unit (mAU) per mg of extracted proteins under the assay conditions. Statistical analysis. Data were analyzed using the JMP software package (version 3.2.6; SAS Institute, Inc., Cary, NC). Mean values for sclerotial and conidial formation affected by different conidium concentrations or strains were compared using the least significant difference and the Tukey-Kramer multiplecomparison test (P ⱕ 0.05). Significantly different mean values are indicated below.

VOL. 74, 2008


5677

FIG. 1. Sclerotial and conidial density-dependent phenomena of A. flavus. (A) Conidia of wild-type strain NRRL 3357 (101, 102, 103, 104, 105, and 106 conidia) were overlaid on plates containing 10 ml of GMM plus 2% sorbitol and incubated in the dark at 29°C for 7 days as described in Materials and Methods. Sclerotial production was determined gravimetrically (mg [dry weight] per plate), and conidial production was determined by counting conidia with a microscope. The values are means of four replicates, and tests were repeated three times in order to examine reproducibility. Spore suspensions were freshly prepared for each test. Different letters (a, b, and c for conidia and A, B, and C for sclerotia) indicate that values are statistically significantly different (P ⬍ 0.05). Only data for 102 to 106 conidia are shown. (B) Images showing the cell density effect on sclerotial production after 1 week of incubation.

Nucleotide sequence accession number. Nucleotide sequence data reported here have been deposited in the GenBank database under accession number EU486993.

RESULTS Cell density affects conidial and sclerotial production. To determine if the initial conidial population size affects development in A. flavus, 10, 102, 103, 104, 105, 106, and 107 conidia were inoculated onto GMM containing 2% sorbitol and incubated in the dark at 29°C for 7 days. Figure 1A and B show that there was an inverse relationship between conidial formation and sclerotial formation at 7 days, where sclerotial numbers (measured by weight) decreased and conidial numbers in-

creased with increasing cell density. The sclerotial weights were statistically different when cultures grown using low cell densities (102 and 103 conidia/plate) were compared with cultures grown using high cell densities (105 and 106 conidia/ plate); this was also true for the numbers of conidia. Similar phenomena were observed when plates were incubated in the light (data not shown). Extract of spent medium from low-cell-density cultures stimulates sclerotial formation, while extract of spent medium from high-cell-density cultures increases conidiation. To examine the possibility that a quorum-sensing mechanism drives the observed sclerotium-conidium balance, we next examined if spent medium from a low-cell-density culture exhibiting high

5678


APPL. ENVIRON. MICROBIOL.

FIG. 2. Influence of extracts prepared from spent medium from high-cell-density (107 cells/plate) and low-cell-density (103 cells/plate) cultures on density-dependent phenomena. Conidia of wild-type NRRL 3357 (102, 103, 104, 105, and 106 conidia) were overlaid on 10 ml of GMM agar plus 2% sorbitol treated with low-cell-density (103 cells/plate) or high-cell-density (107 cells/plate) extracts and incubated in the dark at 29°C for 7 days. Extracts were diluted 1:5, 1:10, or 1:100 in ethanol, as was the medium control extract. (A) Sclerotial formation at high cell densities (106 and 107 conidia/plate) for cultures incubated with extract from plates containing 103 cells/plate (top panel) or with extract from plates containing 107 cells/plate (bottom panel) for three different dilutions of extract, as indicated. (B) Sclerotium formation is stimulated by adding extract from plates containing 103 cells/plate. The values are means of four replicates. Distances greater than the vertical line between symbols for each conidial concentration are significant (P ⱕ 0.05) as determined by the Tukey-Kramer multiple-comparison test. The data are results from two independent experiments. (C) Conidial number is decreased by extract from plates containing 103 cells/plate at a 1:5 dilution and is increased by extract from plates containing 107 cells/plate at a 1:100 dilution. The conidial test was performed by using four replicates, and the experiment was repeated twice to examine reproducibility. For each extract different letters indicate that the values are significantly different (P ⱕ 0.05) as determined by the Tukey-Kramer multiple-comparison test.

sclerotial production (103 conidia/plate) or from a high-density culture exhibiting a high level of conidial production (107 conidia/plate) could affect development. Extracts were added to cultures grown from inocula containing 102 to 107 conidia/ plate. Addition of the 103-conidia/plate extract had no effect on low-density cultures but eliminated the low- or no-sclerotium phenotype typical of high cell densities (Fig. 2A and B). This was true for all dilutions of the 103-conidia/plate extract (1:5, 1:10, and 1:100), all of which significantly increased sclerotial production in 106- and 107-conidia/plate cultures. In contrast, the 107-conidia/plate extract had no effect on sclerotium formation (Fig. 2A and B). The two extracts had opposite effects on conidiation of 106and 107-conidia/plate cultures (other cell densities were not examined). A 1:5 dilution of the 103-conidia/plate extract sig-

nificantly decreased conidiation in 106- and 107-conidia/plate cultures compared to the control (Fig. 2C). In contrast, a 1:100 dilution of the 107-conidia/plate extract significantly increased the number of conidia (Fig. 2C). These observations suggest that low-density cultures contain a factor(s) that induces sclerotium formation or, alternately, inhibits conidiation, while high-cell-density cultures produce a factor(s) that induces conidiation. Sclerotial density dependence is affected by linoleic acid. Because previous studies showed that polyunsaturated fatty acids, particularly linoleic acid, had a stimulatory effect on sclerotial development at a density of 105 conidia/plate (7), we were interested in determining if this 18:2 fatty acid affected the density-dependent development shown in Fig. 1. Filter disks soaked with 0.1 and 1 mg of linoleic acid were added to

VOL. 74, 2008

cultures, and the production of sclerotia was assessed after 6 days. No effects were observed at either low cell densities (10 to 103 conidia/plate) or high cell densities (106 and 107 conidia/ plate), but addition of 1 mg linoleic acid resulted in a significant increase in sclerotium weight at intermediate cell densities (104 and 105 conidia/plate) compared to the control (Fig. 3A and B). Sclerotial weight increased 1.4-fold for the 104-conidia/ plate cell population and 1.7-fold for the 105-conidia/plate cell population. These changes effectively shifted the density-dependent curve toward increased sclerotial production at intermediate cell densities. To assess if other fatty acids that differ in saturation, chain length, and the position of double bonds affect the densitydependent development similar to linoleic acid, filter disks containing linoleic, oleic (18:1), linolenic (18:2), and arachidonic (20:4) acids were added to cultures grown at densities of 104 and 105 conidia/plate. In contrast to linoleic acid, none of these fatty acids stimulated sclerotial formation, and oleic and linolenic acids had the opposite effect at a density of 105 conidia/plate (that is, they depressed sclerotial formation) (Fig. 3C). All studies were compared to the ethanol treatment, as ethanol by itself stimulated sclerotial production at a density of 105 conidia/plate (Fig. 3C). Identification of A. flavus LOX and sequence homology. BLAST searches of the A. flavus genome (http://www.aspergillusflavus.org /genomics) with the A. fumigatus putative arachidonate 5-lipoxygenase (accession no. XP_746463) revealed the presence of only one gene encoding a LOX homolog, which was designated Aflox. The Aflox gene is located on chromosome 3 (GenBank accession no. EU486993). Based on the predicted annotation, Aflox encodes a 574-amino-acid protein after a four-intron splicing event of mRNA. When the predicted amino acid sequence of AfLOX was subjected to a BLASTP search, the program reported homology with the consensus sequences of the LOX family (Pfam 00305; “LOX motif,” His-X4-His-X4His-X17-His-X8-His; http://www.ncbi.nlm.nih.gov/Structure/cdd/). The iron atom in LOX is bound by four ligands, three of which are histidine residues. Six histidines are conserved in all LOX sequences, and five of them are clustered in a stretch of 40 amino acids and are important for LOX activity. Residues that act as iron coordination ligands in plant (48) and mammalian (17, 23) LOXs are very well conserved and are present in AfLOX. When the AfLOX sequence was used as a query with the Protein Data Base, the highest scores were the scores for predicted and known fungal LOXs, followed by mammalian and plant LOXs. The ClustalW software program was used to align A. flavus LOX with several fungal, plant, and mammalian LOXs, and a tree was created by using the TreeView software program (see Fig. S1A in the supplemental material). The predicted AfLOX exhibited 75 and 77% sequence identity with arachidonate 5-lipoxygenase of A. fumigatus (accession no. XP_746463) and N. fischeri LOX (accession no. XP_ 001262545), respectively, and 46% identity with manganese LOX of G. graminis (accession no. AAK81883). Unlike the G. graminis manganese LOX (49), AfLOX is predicted to have the more common nonheme iron at its catalytic center. The homology of AfLOX to the LOX family included homology in the region that contains two His residues found in ␣ helix 9 in the characteristic 30-amino-acid sequence WLLAK-X15-H-X4H-X3-E (27). However, the sequence of AfLOX contained


5679

only 29 amino acids, WLFAK-X14-H(310)-X4-H(315)-X3-E, as did the N. fischeri (accession no. XP_001262545) and A. fumigatus (accession no. XP_746463) LOXs (see Fig. S1B in the supplemental material). Residues important in plant and mammalian LOX catalysis are conserved in AfLOX and other putative fungal LOXs (see Fig. S1B in the supplemental material). A number of residues were completely conserved in fungal LOXs. Identification of A. flavus ⌬lox mutants. The LOX gene of A. flavus was deleted by homologous recombination by replacing Aflox with A. parasiticus pyrG in a pyrG1 mutant (NRRL 3357.5), as shown in Fig. 4A. A PCR screen of 100 transformants for integration of pyrG at the Aflox locus was employed. Transformants showing replacement of Aflox by pyrG by PCR were then confirmed by Southern blot analysis, in which expected hybridization band patterns for a gene replacement event were identified in several transformants (Fig. 4B). Three transformants with Aflox deleted that had identical genotypes as determined by Southern analysis, TSHB2.32, TSHB2.39, and TSHB2.40 (Fig. 4B), were used for physiological analysis. The following two strains were used as controls: the prototroph A. flavus wild-type strain NRRL 3357 and TJSpara17, which was obtained by transforming NRRL 3357.5 to prototrophy with pJW24 containing the A. parasiticus pyrG gene alone. Extensive physiological testing showed that the two control strains have the same phenotype, as demonstrated below (also data not shown). Complementation of Aflox in the ⌬lox background was confirmed by Southern analysis to determine plasmid copy number. Strains TSHB3.5C and TSHB.2(3)C harbored at least two copies of Aflox (Fig. 4B) (extra bands were observed for both these strains), while TSHB3.1C contained a single-copy insertion. These strains were examined to determine if copy number had an effect on Aflox expression, LOX activity, and fungal development. Density-dependent development is minimized in ⌬lox strains. Because inhibition of LOX activity by chemical inhibitors has been shown to affect fungal morphological shifts (28), including those in Aspergillus (A. M. Calvo and N. P. Keller, unpublished data), we thought that it was possible that the densitydependent development demonstrated for the wild type could be affected in lox mutants. The relative ability of ⌬lox strains TSHB2.32, TSHB2.39, and TSHB2.40 to form sclerotia and conidia was determined using the conditions described above for Fig. 1. Conidial production and sclerotial production by the A. flavus wild-type NRRL 3357 and TJSpara17 control strains and complemented strains TSHB3.1C and TSHB3.5C were similarly regulated by cell density. However, the pattern of development was greatly skewed in the ⌬lox strains. The most obvious phenotype was the greatly increased production of sclerotia in all ⌬lox strains at high cell densities (Fig. 5A and 5B). Unlike the the sclerotial production by the wild-type controls, the sclerotial production was only minimally reduced at high inoculum concentrations (105 to 107 conidia/plate). This profile was nearly identical to the profile observed after addition of extract from spent medium from 103-conidia/plate cultures (Fig. 2B). Complemented strains TSHB3.1C and TSHB3.5C had a wild-type phenotype, thus confirming that the effects on sclerotial formation were due to deletion of the lox gene (Fig. 5B). Conidial production was also examined in these

5680



FIG. 3. Linoleic acid increases sclerotial production at intermediate cell densities, whereas linolenic and oleic acids decrease sclerotial production. (A) Filter disks containing 0.1 and 1 mg of linoleic acid and ethanol as a solvent control were added to cultures which were overlaid with 101, 102, 103, 104, 105, 106, and 107 conidia/plate. Sclerotial production was determined gravimetrically (mg [dry weight] per plate). The values are means of three replicates. Distances greater than the vertical bar between symbols for each concentration are significant (P ⱕ 0.05) as determined by the Tukey-Kramer multiple-comparison test. (B) Linoleic acid effect on sclerotial production with 104 conidia/plate. The images were obtained 1 week after inoculation. The upper row shows plates containing ethanol solvent control disks, and the lower row shows plates containing 1 mg linoleic acid dissolved in ethanol on disks. (C) Comparison of the effects of arachidonic, linolenic, linoleic, and oleic acids on sclerotial production at intermediate cell densities. Filter disks containing 1 mg of each of the fatty acids and filter disks containing ethanol as a solvent control were added to cultures which were overlaid with 104 or 105 conidia/plate. Sclerotial production was determined gravimetrically (mg [dry weight] per plate). The values are means of three replicates. Different letters (A and B for 104 conidia/plate and a, b, and c for 105 conidia/plate) indicate statistically significant differences (P ⬍ 0.05) for the cell densities as determined by the Tukey-Kramer multiple-comparison test. The entire experiment was repeated twice, and similar results were obtained.

VOL. 74, 2008


5681

FIG. 4. Deletion of the LOX gene in A. flavus. (A) Diagram of the Aflox coding sequence and how it was replaced with the A. parasiticus pyrG selectable marker by homologous recombination to generate the ⌬lox mutants. (B) Southern analysis of wild-type strain NRRL 3357, ⌬lox mutants TSHB2.39 and TSHB2.40, and complemented strains TSHB3.5C, TSHB3.1C, and TSHB3.2(3)C. Genomic DNA was digested with ApaI, NcoI, and PstI to determine Aflox deletion in the ⌬lox strains or plasmid copy number in the complemented strains. The expected hybridization band patterns for ApaI digestion were a single 3.9-kb band for the wild-type strain, 3.1- and 2.2-kb bands for the ⌬lox strains, and an additional upper band at ⱖ9.5-kb for complemented strains TSHB3.1C and TSHB3.2(3)C. The expected hybridization band patterns for NcoI digestion were 4.1-, 1.3-, and 0.77-kb bands for the wild-type strain, 4.5- and 1.2-kb bands for the ⌬lox strains, and an additional ca. 2.4-kb band for the complemented strains. The expected hybridization band patterns for PstI were 1.9- and 3.3-kb bands for the wild-type strain, 2.4- and 3.2-kb bands for the ⌬lox strains, and additional 5.1- and 4.3-kb bands for an Aflox copy insertion in the complemented strains. Note the extra copies of Aflox in TSHB3.2(3)C and TSHB3.5C (indicated by asterisks). A 3.9-kb PCR product containing Aflox and the 5⬘ and 3⬘ flanking regions was used as a lox probe (indicated by the arrow in panel A).

strains. Again similar to the results obtained for 103-conidia/ plate extracts, the ⌬lox strains produced significantly fewer (P ⱕ 0.05) conidia at high cell densities (ⱖ105 conidia/plate). Complemented strains TSHB3.1C (Fig. 5C) and TSHB3.5C (data not shown) showed wild-type growth. Expression of Aflox. Aflox expression was examined in A. flavus wild-type strain NRRL 3357, ⌬lox strain TSHB2.39, and complemented strains TSHB3.1C and TSHB3.5C. Aflox is a low-expression gene under the conditions used as expression was not detected in wild-type strain NRRL 3357 (Fig. 6A, lanes a and b, and Fig. 6B, lanes a, b, g, and h) or, as expected, in the ⌬lox strain (Fig. 6A, lanes c and d, and Fig. 6B, lanes c, d, k, and l). However, expression of Aflox was observed in the complemented strains. While complemented strain TSHB3.1C (harboring a single-copy insertion) expressed a clear transcript (Fig. 6A, lanes e and f), complemented strain TSHB3.5C (harboring at least two copies of Aflox and likely more) produced a greatly increased transcript and also produced an additional smaller band (Fig. 6B, lanes e, f, k, and l and data not shown). This small band was not observed for single-copy strain TSHB3.1C.

LOX activity and fatty acid analysis. LOX enzymes can utilize many fatty acids (39), and linoleic acid is one of the preferred substrates (33, 43, 49), as well as a common lipid constituent of Aspergillus (6). Therefore, cell fractions of wild-type strain NRRL 3357, ⌬lox strain TSHB2.39, and the complemented strains were examined for linoleic acid LOX activity. Table 2 shows that the LOX activity of the ⌬lox strain (6.5 ⫾ 6.8 1 mAU) was less than that of either the wild-type strain or the complemented strains. The minimal LOX activity detected in ⌬lox strain TSHB2.39 likely resulted from trace activity and may have represented background or nonspecific peroxidation. There was no significant difference in LOX activity between the wild-type strain (97.1 ⫾ 8.7 mAU) and the complemented strains. However, the two multicopy complemented strains, TSHB3.5C (82.9 ⫾ 7.5 mAU) and TSHB3.2(3)C (85.6 ⫾ 8.6 mAU), had slightly lower LOX activity than the single-copy complemented strain, TSHB3.1C (113.2 ⫾ 8.1 mAU). This may have reflected the putative degradation of the transcript in multicopy strains (Fig. 6B). Because oxylipin mutants of A. nidulans have altered fatty

5682



FIG. 5. Effect of Aflox deletion on sclerotial and conidial density-dependent regulation. Conidia of wild-type NRRL 3357 and TJSPara17 control strains, ⌬lox strains TSHB2.32, TSHB2.39, and TSHB2.40, and complemented strains TSHB3.5C and TSHB3.1C (102, 103, 104, 105, 106, and 107 conidia) were overlaid on 10 ml of GMM plus 2% sorbitol and incubated in the dark at 29°C for 7 days. (A) Sclerotial formation after 1 week for densities of 105 to 107 conidia/plate. (B) Sclerotium production increased at high population densities in ⌬lox strains. Each data point is the mean of three replicates. Distances greater than the vertical bar at each concentration are significant (P ⱕ 0.05) as determined by the Tukey-Kramer multiple-comparison test. (C) Conidial production decreased at high population densities in ⌬lox strains. Conidial production at densities of 102, 104, 105, and 106 cells/plate was determined after 3 days. For each of the fungal strains, the conidial test was performed by using five replicates. Different letters indicate that values are significantly different (P ⱕ 0.05) as determined by the Tukey-Kramer multiple-comparison test.

acid contents (52), we next examined any possible effects of the ⌬lox mutation on fatty acid biosynthesis. The mycelial fatty acid contents of the ⌬lox and TSHB3.5C strains were compared to that of wild-type strain NRRL 3357 using gas chromatography analysis of mycelia grown under dark conditions at 29°C for 5 days (Table 3). Whereas deletion of Aflox did not greatly alter the overall fatty acid profile for saturated and unsaturated fatty acids, there were some small effects on indi-

FIG. 6. Aflox expression studies. (A) Aflox expression in liquid cultures of wild-type strain NRRL 3357, ⌬lox strain TSHB2.39, and complemented strain TSHB3.1C assessed as described in Materials and Methods. Three separate repetitions of the experiment yielded similar results. The upper panel was probed with a 1.07-kb internal PCR fragment of Aflox. The middle panel was probed with actin, and rRNA stained with ethidium bromide to indicate RNA loading is shown in lower panel. (B) Aflox is degraded in multicopy strain TSHB3.5C. Liquid cultures of strain NRRL 3357, ⌬lox strain TSHB2.39, and strain TSHB3.5C were transferred to solid medium to induce conidiophore formation as described in the text. Aflox expression in the complemented strain TSHB3.5C was accompanied by a small transcript(s), as indicated by the lower arrow at 12 h (lanes a to f) and 24 h (lanes g to l) after the shift to solid culture. Two separate repetitions of the experiment yielded similar results. The upper panel was probed with a 1.07-kb internal PCR fragment of Aflox, and rRNA stained with ethidium bromide to indicate RNA loading is shown in the lower panel.

TABLE 2. LOX activities of mycelial extracts from A. flavus wild-type strain NRRL 3357, ⌬lox strain TSHB2.39, and complemented strains TSHB3.5C, TSHB3.1C, and TSHB3.2(3)C Samplea

LOX enzymatic activity (U)b

NRRL 3357 (wild type) ........................................................ 97.1 ⫾ 8.7 TSHB2.39 (⌬lox) .................................................................... 6.5 ⫾ 6.8 TSHB3.5C (complemented lox) ........................................... 82.9 ⫾ 7.5 TSHB3.1C (complemented lox) ...........................................113.2 ⫾ 8.1 TSHB3.2(3)C (complemented lox) ...................................... 85.6 ⫾ 8.6 a Total proteins were extracted from NRRL 3357, TSHB2.39, TSHB3.5C, TSHB3.1C, and TSHB3.2(3)C, and LOX activity was assayed as described in the text. b One unit of enzymatic specific activity was defined a change in absorbance at 234 nm of 1 mAU per mg of extracted proteins under the assay conditions used. The data are the means ⫾ standard errors from three experiments.

VOL. 74, 2008


5683

TABLE 3. Fatty acid composition of mycelia of A. flavus wild-type strain NRRL 3357, ⌬lox strain TSHB2.39, and complemented strain TSHB3.5Ca % ofb: Strain

NRRL 3357 TSHB2.39 TSHB3.5C

14:0

16:0

16:1

18:0

18:1

18:2

18:3

2.3 ⫾ 0.7 0.6 ⫾ 0.1 0.4 ⫾ 0.0

17.3 ⫾ 0.4 19.3 ⫾ 0.2 18.3 ⫾ 1.7

2.7 ⫾ 1.0 1.3 ⫾ 0.1 0.6 ⫾ 0.0

10.3 ⫾ 0.9 9.3 ⫾ 0.4 11.7 ⫾ 1.0

28.0 ⫾ 0.2 30.4 ⫾ 0.1 29.1 ⫾ 2.0

38.0 ⫾ 2.6 38.7 ⫾ 0.7 38.5 ⫾ 1.0

1.3 ⫾ 0.0 0.5 ⫾ 0.1 1.4 ⫾ 0.1

a The analysis was carried out with 5-day-old mycelia grown in liquid GMM containing 2% sorbitol at a concentration of 105 conidia/plate under stationary conditions at 29°C in the dark. b The values are the means ⫾ standard errors for three replicates and are the weight percentages of fatty acid methyl esters based on the lyophilized weight of the mycelia.

vidual fatty acids. Some of the fatty acid differences were similar in the deletion strain and the complemented strain and likely are not significant for the developmental processes described here. Unique to the deletion strain were slight increases in the levels of palmitic acid (16:0) and oleic acid (18:1) and a decrease in the level of linolenic acid (18:3). Similar to other Aspergillus species, linoleic acid, whose levels were equivalent in all strains, is the major constituent of A. flavus fatty acids and comprises 38% of the total fatty acid content in A. flavus mycelia. DISCUSSION In this study cell density was found to have a profound effect on morphological transitions in A. flavus; low cell densities yielded primarily sclerotial cultures, and high cell densities yielded conidial cultures. Cell density alteration of development has been observed in other fungi, including C. albicans, C. ulmi, Histoplasma capsulatum, and S. cerevisiae (25, 26, 28, 32). These four organisms are dimorphic yeasts, and much of the literature describes cell density effects on the yeast-tohypha switch; e.g., C. ulmi develops as a budding yeast at high densities and as filaments at low densities (26). Studies of C. albicans and S. cerevisiae have identified quorum-sensing molecules governing this switch. In the former organism farnesol blocks the filament-to-yeast transition, whereas tyrosol promotes the yeast-to-filament switch; all of this is density dependent (11, 44). Two aromatic alcohols, phenylethanol and tryptophol, induce filamentation in S. cerevisiae (10). Although a molecule that conditions the C. ulmi switch has not been identified, chemical inhibition of LOX activity promotes the mycelium-to-yeast conversion in C. ulmi (28). The latter data suggest that an oxylipin signal is involved in cell density-dependent phenomena in this fungus. Cell density regulates development transitions in A. flavus. While filamentous fungi do not exhibit yeast growth patterns, they do develop several tissue types. Members of the genus Aspergillus display several morphologies, including vegetative hyphal growth, asexual sporulation (conidia produced on conidiophores), and in some species sclerotium or cleistothecium formation. Several studies have indicated that sclerotia or cleistothecia are produced under conditions different than those favoring conidiophore formation. Conidiophore production is induced by light, and sclerotia and cleistothecia are produced more abundantly in the dark (5, 6, 7, 37). In A. flavus, circadian rhythms regulate both sclerotial and conidial formation, and conidial production is separated from sclerotial production

(19; and N. P. Keller and D. Bell-Pedersen, unpublished results). Here we show that A. flavus sclerotial formation and conidial formation are regulated in opposite ways by cell density (Fig. 1). High cell densities result in pure conidial cultures, whereas populations with high sclerotial densities are supported at low cell densities. The balancing of sclerotium and conidium levels was reminiscent of a lipid-mediated cleistothecium-conidium balance discovered for A. nidulans (52) and, to a certain extent, for another strain of A. flavus (7), leading to the suggestion that a quorum-driven mechanism may regulate morphological transitions in Aspergillus spp. (50). Supporting this hypothesis, spent medium from low-cell-density, high-sclerotium-density cultures was found to greatly increase sclerotium formation in high-cell-density cultures normally devoid of sclerotia (Fig. 2). Extracts from high-cell-density cultures did not affect sclerotial production at any cell density but led to a significant increase in the number of conidia (Fig. 2C). We suggest that high-celldensity extracts contain a factor(s) that induces conidiation, in contrast to the low-cell-density extracts, which contain a factor(s) that stimulates sclerotial biogenesis and/or inhibits conidiogenesis. Lipid mediation of morphological transitions. Lipids have been implicated repeatedly as signals that regulate fungal reproductive development and growth, and in pathogenic fungi they are involved in signal communication with host cells (4, 24, 30, 39). Furthermore, in the model fungus A. nidulans the cleistothecium-conidium balance is governed by oxylipins (51, 52), reminiscent of the sclerotium-conidium balance observed here. The accumulating data support the hypothesis that lipid pools may be involved in cell density transitions in Aspergillus. Unsaturated fatty acids in particular have been shown to affect fungal development in several genera. Fatty acids serve as signals to initiate the filamentous growth needed for invading plant tissue in the smut pathogen Ustilago maydis (29), and linoleic acid has been shown to induce perithecia in N. crassa (40). Interestingly, fatty acid oscillations are circadian regulated in N. crassa (45), suggesting that there may be an interaction between lipid signals and the circadian regulation of development in A. flavus observed previously (19). Likewise, Goottwald and Wood (18) reported that some fatty acids inhibit sporulation of Cladosporium caryigenum. In Aspergillus spp. fatty acid stimulation of sporulation and/or sclerotial and cleistothecial production is dependent on chain length and the presence of double bonds (7). Polyunsaturated fatty acids had the greatest effect on morphology; for example, linoleic acid

5684


was shown to induce sclerotial production and decrease conidial production in cultures of A. flavus grown using initial cell densities of 105 cells/plate. Our study showed that the sclerotial stimulation by linoleic acid is limited to intermediate cell densities (Fig. 3). This effect could be a direct result of use of linoleic acid as a carbon source and/or an indirect result through metabolic pathways utilizing this fatty acid. In contrast to linoleic acid, oleic and linolenic acids had inhibitory effects on sclerotium formation. Role of LOX in A. flavus cell density transitions. Unsaturated fatty acids can be metabolized by beta-oxidation pathways (35) or used as substrates by various oxygenases, including LOXs, dioxygenases, and P450 monooxygenases, to generate diverse oxylipins (34, 49). Three key observations led us to examine the possible role of A. flavus LOX in the cell-densitydependent conidium-to-sclerotium switch. First, chemical inhibition of LOX interfered with the dimorphic switch in C. ulmi (28). Second, physiological studies of Aspergillus showed that purified oxylipins could affect morphological development. For example, the plant linoleic LOX product 9(S)-hydroperoxy-trans-10-cis-12-octadecadienoic acid (9S-HPODE) induced cleistothecial formation in A. nidulans at levels of 0.01 and 0.1 mg but conidial formation at 10-fold-higher concentrations (7). Third, genetic disruptions of oxylipin-generating dioxygenases (ppo genes) in A. nidulans greatly altered the conidium-cleistothecium balance in this species; loss of one dioxygenase yielded a nearly pure conidial culture (52), and loss of two other dioxygenases yielded a predominantly cleistothecial culture (51). Because analysis of the A. flavus genome indicated that there are four dioxygenases (ppoA, ppoB, ppoC, and ppoD) and it would be difficult to disrupt all four genes in this asexual species, we chose to characterize the single, putative LOX gene designated Aflox. Deletion of Aflox eliminated the switch from a sclerotial culture to a conidial culture (Fig. 5). In the wild type and the complemented controls, conidial production significantly increased and sclerotial production decreased at cell densities between 104 and 105 cells/plate. At concentrations of ⱖ106 cell/plate, no sclerotia were produced and conidial production was maximal. In the ⌬Aflox mutants, sclerotial production decreased minimally and conidial production remained low at these cell densities. This phenotype was very similar to that observed when extract from spent medium from a plate containing 103 conidia was added to growing cultures (Fig. 2). This suggests that some AfLOX-derived oxylipin and/or a downstream product(s) signals sclerotium-to-conidium transitions in A. flavus. These observations are reminiscent of the concentration-dependent effect of the oxylipin 9S-HPODE on A. nidulans cleistothecium-to-conidium shifts (7). The LOX activity assay showed that there was a significant reduction in linoleic acid-mediated activity in the ⌬Aflox strain compared to the wild-type and complemented strains, suggesting that linoleic acid is an AfLOX substrate. However, it is well established that oxylipin-generating enzymes (dioxygenase, LOX, and cyclooxygenase) exhibit activity with more than one substrate, and it is possible that AfLOX utilizes other unsaturated fatty acids as substrates. AfLOX shows the highest levels of homology to predicted arachidonate 5-lipoxygenases in other aspergilli, which suggests that arachidonic acid, a fatty acid found at negligible levels in Aspergillus spp., may also be a


substrate for AfLOX. We note that, along with changes in LOX activity, loss of Aflox resulted in slight changes in fatty acid percentages, including a slight increase in the oleic acid content and a decrease in the linolenic acid content. However, the developmental effects observed for the ⌬Aflox strain are more likely due to changes in oxylipin content than to changes in fatty acids, as one would predict that a slight increase in the content of oleic acid (which likely has the greatest effect due to the total percentage of this fatty acid) would reduce sclerotial production (Fig. 3C) rather than increase production, the phenotype of ⌬Aflox. Therefore, we postulate that one or more AfLOX-generated oxylipins may play a role in quorum sensing, as observed from the lack of a morphological shift in the ⌬Aflox strains (Fig. 5). Such AfLOX products, possibly promoting conidial development, would be in opposition to the factor(s) acting in the spent medium from low-cell-density cultures (Fig. 2). The undetectable level of Aflox in the wild-type strain under the conditions tested (Fig. 6) may reflect the low levels of a putative quorum-sensing molecule needed or may simply reflect the cultural conditions used. Similarly, the level of expression of ppoB, which is involved in the production of oleic acid-derived oxylipin in A. nidulans, was very low despite a profound effect on the conidium-to-cleistothecium switch (52). Interestingly, differences in Aflox expression among the different complemented strains had no effect on conidium-to-sclerotium transitions. As the LOX activities were similar for all of the complemented strains, it is possible that there is some gene-silencing phenomenon, possibly RNA interference activity (20), in multipcopy strains and/or that there is posttranscriptional regulation of this enzyme. These findings support the hypothesis that a quorum-sensing mechanism governs morphological shifts in the filamentous fungus A. flavus. The chemical signals integrate cell density and development by regulating specialized processes such as morphogenesis, perhaps as a means of responding effectively to changes in environmental opportunities. The conidium is the dispersal spore expected to be produced at times of starvation, conditions that may arise with high population densities and cells competing for limited nutrients. The sclerotium is a resistant body, establishing a niche for the producing organism and deterring predators through production of toxic metabolites (53). A sentinel system alerting the fungus to environments favoring “flight” versus “fight” would be evolutionarily beneficial for the organism. ACKNOWLEDGMENTS We thank Gary A. Payne for access to A. flavus genome data and for providing strains, James Scott for critical reading of the manuscript, and Jon P. Woods and Stacey Schultz-Cherry for sharing some of the equipment used in this study. This research was funded by BARD grant FI-384-2006 to S.H.B. and by NSF grant IOB-0544428 (subagreement S060039) to N.P.K. REFERENCES 1. Allison, C., and C. Hughes. 1991. Bacterial swarming: an example of prokaryotic differentiation and multicellular behavior. Sci. Prog. 75:403–422. 2. Axelrod, B., T. M. Cheesebrough, and S. Laakso. 1981. Lipoxygenase from soybean. Methods Enzymol. 71:441–451. 3. Bhatnagar, D., J. W. Carry, K. Ehrlich, J. Yu, and T. E. Cleveland. 2006. Understanding the genetics of regulation of aflatoxin production and Aspergillus flavus development. Mycopathologia 162:155–166. 4. Brodhagen, M., D. I. Tsitsigiannis, E. Hornung, C. Goebel, I. Feussner, and

VOL. 74, 2008

5.

6.

7.

8.

9. 10. 11.

12.

13.

14. 15.

16. 17.

18.

19.

20.

21.

22.

23.

24.

25. 26.

27.

28.

N. P. Keller. 2008. Reciprocal oxylipin-mediated cross-talk in the Aspergillusseed pathosystem. Mol. Microbiol. 67:378–391. Calvo, A. M., J. W. Bok, W. Brooks, and N. P. Keller. 2004. VeA is required for toxin and sclerotial production in Aspergillus parasiticus. Appl. Environ. Microbiol. 70:4733–4739. Calvo, A. M., H. W. Gardner, and N. P. Keller. 2001. Genetic connection between fatty acid metabolism and sporulation in Aspergillus nidulans. J. Biol. Chem. 276:25766–25774. Calvo, A. M., L. L. Hinze, H. W. Gardner, and N. P. Keller. 1999. Sporogenic effect of polyunsaturated fatty acids on development of Aspergillus spp. Appl. Environ. Microbial. 65:3668–3673. Calvo, A. M., R. A. Wilson, J. W. Bok, and N. P. Keller. 2002. Relationship between secondary metabolism and fungal development. Microbiol. Mol. Biol. Rev. 66:447–459. Champe, S. P., and A. A. El-Zayat. 1989. Isolation of a sexual sporulation hormone from Aspergillus nidulans. J. Bacteriol. 171:3982–3988. Chen, H., and G. R. Fink. 2006. Feedback control of morphogenesis in fungi by aromatic alcohols. Genes Dev. 20:1150–1161. Chen, H., M. Fujita, Q. Feng, J. Clardy, and G. R. Fink. 2004. Tyrosol is a quorum-sensing molecule in Candida albicans. Proc. Natl. Acad. Sci. USA 101:5048–5052. Chenna, R., H. Sugawara, T. Koike, R. Lopez, T. J. Gibson, D. G. Higgins, and J. D. Thompson. 2003. Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Res. 31:3497–3500. Diener, U. L., R. J. Cole, T. H. Sanders, G. A. Payne, L. S. Lee, and M. A. Klich. 1987. Epidemiology of aflatoxin formation by Aspergillus flavus. Annu. Rev. Phytopathol. 25:249–270. Erb-Downward, J. R., and G. B. Huffnagle. 2006. Role of oxylipins and other lipid mediators in fungal pathogenesis. Future Microbiol. 1:219–227. Fuqua, W. C., S. C. Winans, and E. P. Greenberg. 1994. Quorum sensing in bacteria: the LuxR-LuxI family of cell density-responsive transcriptional regulators. J. Bacteriol. 176:269–275. Geiser, D. M., W. E. Timberlake, and M. L. Arnold. 1996. Loss of meiosis in Aspergillus. Mol. Biol. Evol. 13:809–817. Gillmor, S. A., A. Villasenor, R. Fletterick, E. Sigal, and M. F. Browner. 1997. The structure of mammalian 15-lipoxygenase reveals similarity to the lipases and the determinants of substrate specificity. Nat. Struct. Biol. 4:1003–1009. Gottwald, T. R., and B. W. Wood. 1984. The effect of fatty acids on growth and sporulation of the pecan scab fungus Cladosporium caryigenum. Mycologia 76:326–331. Greene, A. V., N. P. Keller, H. Haas, and D. Bell-Pedersen. 2003. A circadian oscillator in Aspergillus spp. regulates daily development and gene expression. Eukaryot. Cell 2:231–237. Hammond, T. M., and N. P. Keller. 2005. RNA silencing in Aspergillus nidulans is independent of RNA dependent RNA polymerases. Genetics 169:607–617. He, Z. M., M. S. Price, G. R. Obrian, D. R. Georgianna, and G. A. Payne. 2007. Improved protocols for functional analysis in the pathogenic fungus Aspergillus flavus. BMC Microbiol. 7:104. Hedayati, M. T., A. C. Pasqualotto, P. A. Warn, P. Bowyer, and D. W. Denning. 2007. Aspergillus flavus: human pathogen, allergen and mycotoxin producer. Microbiology 153:1677–1692. Hegg, E. L., and L. Que, Jr. 1997. The 2-His-1-carboxylate facial triad—an emerging structural motif in mononuclear non-heme iron(II) enzymes. Eur. J. Biochem. 250:625–629. Herman, R. P. 1998. Oxylipin production and action in fungi and related organisms, p. 115–132. In A. F. Rowley, H. Kuhn, and T. Schewe (ed.), Eicosanoids and related compounds in plants and animals. Princeton University Press, Princeton, NJ. Hogan, D. A. 2006. Talking to themselves: autoregulation and quorum sensing in fungi. Eukaryot. Cell 5:613–619. Hornby, J. M., S. M. Jacobitz-Kizzier, D. J. McNeel, E. C. Jensen, D. S. Treves, and K. W. Nickerson. 2004. Inoculum size effect in dimorphic fungi: extracellular control of yeast-mycelium dimorphism in Ceratocystis ulmi. Appl. Environ. Microbiol. 70:1356–1359. Hughes, R. K., D. M. Lawson, A. R. Hornostaj, S. A. Fairhurst, and R. Casey. 2001. Mutagenesis and modeling of linoleate-binding to pea seed lipoxygenase. Eur. J. Biochem. 268:1030–1040. Jensen, E. C., C. Ogg, and K. W. Nickerson. 1992. Lipoxygenase inhibitors


29. 30. 31.

32.

33.

34. 35.

36. 37. 38.

39.

40.

41.

42. 43.

44.

45. 46.

47.

48.

49. 50. 51.

52.

53.

5685

shift the yeast/mycelium dimorphism in Ceratocystis ulmi. Appl. Environ. Microbiol. 58:2505–2508. Klose, J., M. M. Moniz de Sa ´, and J. W. Kronstad. 2004. Lipid-induced filamentous growth in Ustilago maydis. Mol. Microbiol. 52:823–835. Kock, J. L., C. J. Strauss, C. H. Pohl, and S. Nigam. 2003. The distribution of 3-hydroxy oxylipins in fungi. Prostaglandins Other Lipid Mediat. 71:85–96. Kolbinger, A., T. Gao, D. Brock, R. Ammann, A. Kisters, J. Kellermann, D. Hatton, R. H. Gomer, and B. Wetterauer. 2005. A cysteine-rich extracellular protein containing a PA14 domain mediates quorum sensing in Dictyostelium discoideum. Eukaryot. Cell 4:991–998. Kugler, S., T. Schurtz Sebghati, L. Groppe Eissenberg, and W. E. Goldman. 2000. Phenotypic variation and intracellular parasitism by Histoplasma capsulatum. Proc. Natl. Acad. Sci. USA 97:8794–8798. Kuribayashi, T., H. Kaise, C. Uno, T. Hara, T. Hayakawa, and T. Joh. 2002. Purification and characterization of lipoxygenase from Pleurotus ostreatus. J. Agric. Food Chem. 50:1247–1253. Liavonchanka, A., and I. Feussner. 2006. Lipoxygenases: occurrence, functions and catalysis. J. Plant Physiol. 163:348–357. Maggio-Hall, L. A., R. A. Wilson, and N. P. Keller. 2005. Fundamental contribution of beta-oxidation to polyketide mycotoxin production in planta. Mol. Plant Microbe Interact. 18:783–793. Miller, B. L., K. Y. Miller, and W. E. Timberlake. 1985. Direct and indirect gene replacements in Aspergillus nidulans. Mol. Cell. Biol. 5:1714–1721. Mooney, J. L., and L. N. Yager. 1990. Light is required for conidiation in Aspergillus nidulans. Genes Dev. 4:1473–1482. Nickerson, K. W., A. L. Atkin, and J. M. Hornby. 2006. Quorum sensing in dimorphic fungi: farnesol and beyond. Appl. Environ. Microbiol. 72:3805– 3813. Noverr, M. C., J. R. Erb-Downward, and G. B. Huffnagle. 2003. Production of eicosanoids and other oxylipins by pathogenic eukaryotic microbes. Clin. Microbiol. Rev. 16:517–533. Nukima, M., T. Sassa, M. Ikeda, and K. Takahashi. 1981. Linoleic acid enhances perithecial production in Neurospora crassa. Agric. Biol. Chem. 45:2371–2373. Oh, K. B., H. Miyazawa, T. Naito, and H. Matsuoka. 2001. Purification and characterization of an autoregulatory substance capable of regulating the morphological transition in Candida albicans. Proc. Natl. Acad. Sci. USA 98:4664–4668. Reference deleted. Perez-Gilabert, M., I. Sanchez-Felipe, A. Morte, and F. Garcia-Carmona. 2005. Kinetic properties of lipoxygenase from desert truffle (Terfezia claveryi Chatin) ascocarps: effect of inhibitors and activators. J. Agric. Food Chem. 53:6140–6145. Ramage, G., S. P. Saville, B. L. Wickes, and J. L. Lopez-Ribot. 2002. Inhibition of Candida albicans biofilm formation by farnesol, a quorum-sensing molecule. Appl. Environ. Microbiol. 68:5459–5463. Roeder, P. E., M. L. Sargent, and S. Brody. 1982. Circadian rhythms in Neurospora crassa: oscillations in fatty acids. Biochemistry 21:4909–4916. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Shimizu, K., and N. P. Keller. 2001. Genetic involvement of a cAMPdependent protein kinase in a G protein signaling pathway regulating morphological and chemical transitions in Aspergillus nidulans. Genetics 157: 591–600. Skrzypczak-Jankun, E., L. M. Amzel, B. A. Kroa, and M. O. Funk, Jr. 1997. Structure of soybean lipoxygenase L3 and a comparison with its L1 isoenzyme. Proteins 29:15–31. Su, C., and E. H. Oliw. 1998. Manganese lipoxygenase. Purification and characterization. J. Biol. Chem. 273:13072–13079. Tsitsigiannis, D. I., and N. P. Keller. 2007. Oxylipins as developmental and host-fungal communication signals. Trends Microbiol. 15:109–118. Tsitsigiannis, D. I., T. M. Kowieski, R. Zarnowski, and N. P. Keller. 2004. Endogenous lipogenic regulators of spore balance in Aspergillus nidulans. Eukaryot. Cell 3:1398–1411. Tsitsigiannis, D. I., T. M. Kowieski, R. Zarnowski, and N. P. Keller. 2005. Three putative oxylipin biosynthetic genes integrate sexual and asexual development in Aspergillus nidulans. Microbiology 151:1809–1821. Wicklow, D. T., B. W. Horn, and R. J. Cole. 1982. Sclerotium production by Aspergillus flavus on corn kernels. Mycologia 74:398–403.