The Role, Interaction and Regulation of the Velvet Regulator VelB in Aspergillus nidulans

The multifunctional regulator VelB physically interacts with other velvet regulators and the resulting complexes govern development and secondary metabolism in the filamentous fungus Aspergillus nidulans. Here, we further characterize VelB’s role in governing asexual development and conidiogenesis in A. nidulans. In asexual spore formation, velB deletion strains show reduced number of conidia, and decreased and delayed mRNA accumulation of the key asexual regulatory genes brlA, abaA, and vosA. Overexpression of velB induces a two-fold increase of asexual spore production compared to wild type. Furthermore, the velB deletion mutant exhibits increased conidial germination rates in the presence of glucose, and rapid germination of conidia in the absence of external carbon sources. In vivo immuno-pull-down analyses reveal that VelB primarily interacts with VosA in both asexual and sexual spores, and VelB and VosA play an inter-dependent role in spore viability, focal trehalose biogenesis and control of conidial germination. Genetic and in vitro studies reveal that AbaA positively regulates velB and vosA mRNA expression during sporogenesis, and directly binds to the promoters of velB and vosA. In summary, VelB acts as a positive regulator of asexual development and regulates spore maturation, focal trehalose biogenesis and germination by interacting with VosA in A. nidulans.


Introduction
Fungal spores are widespread in the environment and have a significant impact on daily human life. Asexual spores are the main reproductive form of many fungi and the primary agent for infecting hosts for many pathogenic fungi [1]. In some Aspergillus, asexual spore formation is closely correlated with the production of toxic secondary metabolites called mycotoxins [2,3]. Among Aspergillus species, Aspergillus nidulans is an excellent model system for studying the regulation of gene expression and the mechanisms of asexual development [4,5].
Asexual development (conidiation) in A. nidulans involves the formation of specialized multicellular structures called conidiophores. Conidiophore formation starts from hyphal cells, which form the mycelium, followed by the sequential formation of vesicles, metulae, phialides and conidia [4,6]. The process of conidiation is genetically regulated and the three genes brlA, abaA and wetA have been proposed to define a central regulatory pathway activating conidiophores formation [4,[7][8][9][10]. After the spore is physically formed, it must undergo the maturation process. During conidiophore maturation, a key process is the focal biogenesis of large amounts (up to 15% of dry weight) of trehalose, a-D-glucopyranosyl-a-D-glucopyranoside, within the spore there-by conferring long term viability [11]. Trehalose, found in a wide variety of bacteria, fungi, plant, and invertebrates, serves as a vital protectant against desiccation and various environmental stresses and as an energy source [12]. Recent studies have identified the novel regulator VosA, which functions in trehalose biosynthesis and conidia maturation in A. nidulans. VosA also exerts negative feedback control of conidiation by down-regulating brlA expression. Moreover, VosA is mainly localized in the nucleus of mature conidia and it contains a potential transcriptional activation domain at the C-terminus. These findings led to the hypothesis that VosA is a transcription factor regulating conidia maturation and the completion of conidiogenesis [13].
The velvet family proteins, including VeA, VelB, VelC and VosA, have defined a novel protein family. These four proteins all contain the velvet domain, and are highly conserved in dimorphic and filamentous fungi [13,14]. A series of recent studies have revealed that the velvet proteins form the multimeric velvet complexes such as VelB/VeA/LaeA, VelB/VosA, and VelB/ VelB [15,16]. The heterotrimeric VelB/VeA/LaeA complex controls sexual development and secondary metabolism in response to light [15]. The nuclear VelB-VosA heterodimer has been proposed to regulate trehalose biogenesis and spore maturation [16].
VelB, a primary component of the velvet complexes, plays a crucial role in sexual development and secondary metabolism in A. nidulans [15,16]. However, additional role and interaction of VelB in the life cycle of conidiogenesis, and temporal regulation of velB expression in the late phase of conidiation in A. nidulans have not been studied. In this study, we further characterize velB by genetic and biochemical approaches. The deletion of velB results in elevated accumulation of brown pigment(s) and decreased production of conidia. Conversely, overexpression of velB results in increased conidiation in air-exposed conditions, suggesting an activating role of VelB in conidiation. In addition, VelB plays a negative role in regulating conidial germination regardless of the presence or absence of an external carbon source. VelB predominantly interacts with VosA in asexual and sexual spores and plays an inter-dependent role in focal trehalose biogenesis, conidial maturation and germination. We further demonstrate that AbaA positively regulates velB and vosA mRNA expression during the mid-to-late phase of conidiation and directly binds to the promoters of these genes. Taken together, we present a model that VelB acts as a positive regulator of conidiation and AbaA activates expression of VelB and VosA in phialides, which then primarily forms the VelB/VosA heterodimer in conidia and controls focal trehalose biosynthesis and conidial germination.

Strains, Media and Culture Conditions
A. nidulans strains used in this study are listed in Table 1. The fungal strains were grown on solid or liquid minimal medium with supplements (simplified as MM) as described previously [13,17] and incubated at 37uC. To determine the number of conidia, wild type (WT) and mutant strains were point inoculated and grown on solid MM at 37uC for 3 to 5 days. The conidia were collected in ddH 2 O from the entire colony and counted using a hemocytometer. For liquid submerged cultures, conidia of WT and mutant strains were inoculated in 50 ml of liquid MM (5610 5 conidia/ml) and incubated at 37uC, 250 rpm. To examine the effects of overexpression (OE) of velB by an ectopic copy of velB under the alcA promoter, all strains were inoculated on solid MM with 1% glucose (MMG) or MM with 100 mM threonine (MMT to induce overexpression of velB) and 0.1% yeast extract at 37uC for 4 days.
For Northern blot analysis, samples were collected as described [13,18,19]. Briefly, for vegetative growth phases, conidia (5610 5 conidia/ml) of WT and mutant strains were inoculated in 100 ml liquid MM in 500 ml flasks and incubated at 37uC. Samples were collected at designated time points of liquid submerged culture, squeeze-dried and stored at 280uC. For sexual and asexual developmental induction, 18 h vegetatively grown mycelia were filtered, washed and transferred to solid MM and the plates were air exposed for asexual developmental induction, or tightly sealed and blocked from light for sexual developmental induction. To examine velB mRNA levels in alcA(p)::brlA, alcA(p)::brlA abaA14 and alcA(p)::abaA strains, three strains were grown in liquid glucose medium (G) at 37uC, 250 rpm for 14 h (designated as time point ''0'') and then transferred into liquid glucose medium (G; noninducing) or liquid threonine medium (T; inducing). Samples were collected at designated time points after transfer. Escherichia coli

Generation of velB Mutants
The oligonucleotides used in this study are listed in Table 2. For the deletion of velB, DJ-PCR method was used [20]. Both flanking regions of velB were amplified using the primer pairs OM-N131;OMN135 and OMN132;OMN136 using A. nidulans FGSC4 genomic DNA as a template. The A. fumigatus pyrG marker was PCR-amplified from A. fumigatus AF293 genomic DNA with the primer pair OJH84;OJH85. The final PCR constructs for the velB deletion were amplified with OMN133;OMN134. The deletion cassettes were introduced into RJMP1.59 protoplasts generated by the Vinoflow FCE lysing enzyme (Novozymes) [21]. To generate the double deletion mutants, 59 and 39 flanking regions of vosA (OMN54;OHS184 and OMN55;OHS185) and ganB (OH-S170;OHS174 and OHS171;OHS175) were amplified. The pyroA marker was amplified from FGSC4 genomic DNA with the primer pair ONK395;ONK396. After the fusion by DJ-PCR, vosA and ganB deletion constructs were amplified using OMN58;OMN59 and OHS172;OHS173, respectively, and introduced into RNI 18.3. Multiple (at least three) double mutants were isolated and confirmed by PCR followed by restricted enzyme digestion in each case.
To generate the alcA(p)::velB fusion construct, the velB ORF derived from genomic DNA was PCR amplified using the primer pair OMN196;OHS141. The PCR product was then double digested with BamHI and HindIII and cloned into pHS3, which contains the A. nidulans alcA promoter and the trpC terminator (this study). The resulting plasmid pHSN6 was then introduced into RJMP1.59 or FGSC 33. The velB overexpression strains among the transformants were screened by Northern blot analysis using a velB ORF probe followed by PCR confirmation. RHS2.1, 2.2, 2.3 and 2.4 were isolated from the sexual cross between THS7.1 and THS18.1.
To complement DvelB, the WT velB gene region including its predicted promoter was amplified with the primer pair OH-S140;OHS141, digested with EcoRI and HindIII and cloned into pHS13. This cloning vector contains the LpyroA gene [22], a 3xFLAG tag and the trpC terminator [23]. The resulting plasmid pHSN31 was then introduced into the recipient DvelB strain THS16.1, where preferentially a single copy velB + gets inserted into the pyroA locus, to give rise to THS20.1.

Nucleic Acid Manipulation
To isolate genomic DNA, about 10 6 conidia of WT and mutant strains were inoculated in 2 ml liquid MMG +0.5% yeast extract, and stationary cultured at 37uC for 24 h. The mycelial mat was collected, squeeze-dried, and genomic DNA was isolated as described [20]. Total RNA isolation and Northern blot analyses were carried out as previously described [18,20,24]. The DNA probes were prepared by PCR-amplification of the coding regions of individual genes with appropriate oligonucleotide pairs using FGSC4 genomic DNA as a template (Table 2).

Spore Viability Test
Spore viability was determined as described [13]. Two-day old conidia (10 5 per plate) of WT and mutant strains were spread onto solid MM and incubated at 37uC. After 2,10 days, the conidia were collected and counted using a hemocytometer. Approximately 200 conidia were spread onto solid MM and incubated for 2 days at 37uC in triplicates. Survival rates were calculated as a ratio of the number of viable colonies to the number of spores inoculated.

Spore Trehalose Assay
Trehalose was extracted from conidia and analyzed as described [13,25]. Two-day old conidia (2610 8 ) were collected and washed with ddH 2 O. Conidia were resuspended in 200 ml of ddH 2 O and incubated at 95uC for 20 min, and then the supernatant was collected by centrifugation. The supernatant was mixed with an equal volume of 0.2 M sodium citrate (pH5.5) and further incubated at 37uC for 8 h with or without (control) 3 mU of trehalase (Sigma), which hydrolyzes trehalose to glucose. The amount of glucose generated by trehalose was assayed with a glucose assay kit (Sigma), and converted into the trehalose amount per conidium (triplicates). Each sample not treated with trehalase served as a negative control.

Germination of Conidia
Germination rates were measured as previously described [26]. The conidia of WT and mutants were spread onto solid MM containing glucose or no carbon source and incubated at 37uC. Levels of germination (isotropic growth and germ tube formation) were examined every 1 h after inoculation under a microscope.

Eletrophoretic Mobility Shift Assay (EMSA)
The cDNA fragment of abaA encoding amino acids 1-588 with the ATTA domain was cloned into pGEX-5X-1 (GE Healthcare).  The resulting plasmid (pHSE3) was introduced into E. coli BL21 (DE3) (Stratagene). The GST fusion protein expression and purification was carried out following the manufacturer's instruction. For concentration and buffer exchange, Amicon Ultra Centrifilter Unit (Milllipore) was used. BCA Protein Assay Kit (Pierce) was used to estimate protein concentration. EMSA was carried out as described [27][28][29]. Probes for EMSA were generated by annealing two single-stranded reverse-complementary oligonucleotides. The ds oligonucleotides were labeled with c-32 P-ATP (PerkinElmer) using T4 polynucleotide kinase (Promega). After labeling, the probes were purified by Sephadex G-25 spin columns. Binding reactions were performed in a 24 ml reaction volume containing 6% glycerol, 12  TBE running buffer at 100 V for 1 h. The gel was dried down under vacuum to three layers Whatman 3 MM filter paper. Autoradiography was performed at 280uC with Kodak XAR film.

Purification of in vivo VelB Interacting Proteins
For sample preparation, conidia and ascospores from two and seven day grown colonies, respectively, of WT and THS 20.1 (DvelB; velB::3xFLAG) strains were collected and suspended in cold PBS with 0.02% Triton X-100. The spores were sonicated for 300 sec (50% pulse) on ice to remove rodlets, and centrifuged for 10 min at 3,000 rpm and the supernatants were removed. For preparation of the protein samples from the mycelium, conidia (5610 5 conidia/ml) of WT and THS 20.1 strains were inoculated in 500 ml of liquid MM and incubated for 18 h at 37uC. Mycelial samples were collected, squeeze dried and stored at 280uC. Prepared spores or mycelial samples were broken by a mini-bead beater for 2 cycles (1 min homogenization with 10 min sitting on ice) and centrifuged in a microcentrifuge for 15 min at 15,000 rpm at 4uC. The supernatant was incubated with anti-FLAG M2 affinity gel (A2220, Sigma). The agarose beads were collected by centrifugation and washed three times. The sample was boiled for 5 min and separated by SDS-PAGE and visualized by Coomassie Blue. Each stained band was cut out and subjected to mass spectrometry.

Nano-LC-ESI-MS/MS and Protein Identifications
All MS/MS experiments for peptide identification were performed using a nano-LC-MS system consisting of an Ultimate HPLC system (Dionex) and Q-TOF mass spectrometer (Micromass, Manchester, UK) equipped with a nano-ESI source, as described previously [30]. Briefly, for each sample, 10 ml was loaded by the autosampler onto a C18 trap column (id 300 mm,  velB::3xFLAG) strain. Arrows indicate the antibody fragments. After staining, the indicated regions were excised and proteins were identified using a length 5 mm, particle size 5 mm; LC Packings) for desalting and concentrating at a flow rate of 20 ml/min. Then the trapped peptides were separated on a 100-mm homemade microcapillary column composed of C18 (Aqua; particle size, 5 mm) packed into 75-mm silica tubing. The mobile phases, A and B, were composed of 0 and 80% acetonitrile, respectively, each containing 0.1% formic acid. The gradient began with 5% B for 15 min; ramped to 20% B over 3 min, to 60% over 45 min, and to 95% over 2 min; and remained at 95% B for another 7 min. The column was equilibrated with 5% B for 10 min before the next run. The voltage applied to produce an electrospray was 2.5 kV. Argon was introduced as a collision gas at a pressure of 10 psi. The three most abundant MS ions were selected by data-dependent peak selection. The previously fragmented ions were excluded for 60 s. The proteins were identified by searching fungi subset (219981 entries) of the National Center for Biotechnology Information (NCBI) protein databases (20100312) using the MASCOT 2.0 search algorithm (Matrix Science). The general parameters for a search were considered to allow a maximum of one missed cleavage, the modifications of N-terminal Gln to pyroGlu, oxidation of methionine, acetylation of protein N terminus, carbamidomethylation of cysteine, and acrylamide modified cysteine. A peptide charge state of +2 or 3, and peptide/fragment mass tolerance of 60.5 Da were used for the MS/MS ion search. Probability based MASCOT scores were estimated by comparison of search results against estimated random match population and reported as 210*Log(P) where P is the absolute probability. The significance threshold was set at p,0.05.

Microscopy
The colony photographs were taken by using a Sony digital camera (DSC-F828). Photomicrographs were taken using a Zeiss M 2 BIO microscope equipped with AxioCam and AxioVision digital imaging software (Zeiss).

Statistical Analysis
Statistical differences between WT and mutant strains were evaluated with student's unpaired t-test (2-tailed). Mean 6 SD are shown. P values ,0.05 were considered significant.

Developmental Defects Caused by the Lack of VelB
To further investigate the roles of velB, we generated velB deletion (DvelB) and complemented strains and compared their phenotypes. As shown in Fig. 1A, DvelB strains produced highly elevated levels of brown pigment(s). In the submerged culture, dark brown pigment(s) accumulated in the DvelB hyphae and was detected in liquid medium (data not shown). The velB deletion mutant also exhibited reduced and delayed conidiation when point inoculated on solid medium (Fig. 1A). When 10 5 conidia were inoculated onto solid medium and incubated for 2 days, the DvelB mutant produced a reduced number of conidia (Fig. 1C); about a tenfold decrease in the yield of conidia compared to WT (Fig. 1B,  1D). Essentially, the same results were observed in the velB deletion strains with the veA1 allele (data not shown).
To correlate phenotypic changes caused by deletion of velB with the molecular events, we examined the mRNA levels of brlA, abaA and vosA in WT and DvelB strains grown under conditions that induce asexual development (Fig. 1E). In WT, accumulation of brlA mRNA was detectable at 6 h post developmental induction and decreased after 24 h. In the DvelB mutant, however, brlA mRNA started to accumulate at 12 h and stayed at high levels even after 24 h. Levels of abaA mRNA decreased and delayed in the DvelB mutant compared to WT, too. While the short transcript of vosA started to accumulate at 24 h and the long transcript of vosA accumulated at 48 h in WT, neither long or short transcripts accumulated at 24 h and 48 h after induction of development in the DvelB mutant (Fig. 1F). These results indicate that VelB is necessary for the proper progression of conidiation.
Previous study reported that the DvelB mutant failed to form sexual fruiting bodies in the light and dark conditions [31]. We examined whether the deletion of velB affected the expression patterns of key sexual developmental genes including mutA, nosA, rosA and nsdD in WT and DvelB strains grown under the conditions that preferentially induce sexual development (Fig. 1G, 1H) [32][33][34][35]. As shown in Fig. 1H, the deletion of velB caused about 24 h delayed (and reduced) accumulation of mutA (a-1,3-glucanase) that is associated with Hülle cells [32]. For rosA (repressor of sexual development; [33]) transcript, WT showed low level accumulation of its mRNA at 0,12 h, which then decreased after 24 h post developmental induction. However, the DvelB mutant continued to exhibit (high level) accumulation of rosA mRNA even at 24 h post developmental induction, i.e., ,12 h delay in turning off rosA expression compared to WT. During early sexual developmental induction (0,12 h), levels of nosA (number of sexual spores; [34]) mRNA slightly decreased in the DvelB mutant compared to WT. Levels of nsdD (never in sexual development; [35]) mRNA were not distinctly different between WT and DvelB strains. These results indicate that VelB is required for proper and timely expression of mutA/rosA/nosA in A. nidulans, which is critical for the development of Hülle cells and cleistothecia.

VelB is Conditionally Sufficient to Enhance Conidiation
We then generated the velB overexpression mutant by fusing the velB ORF (open reading frame) with the inducible alcA promoter [36]. Under non-inducing conditions, there were no differences between WT and overexpression of velB. When point inoculated under inducing conditions, overexpression of velB with the veA + or veA1 allele induced a two-fold increase of asexual spore production compared to WT ( Fig. 2A, 2B). We then tested whether VelB has a direct activating potential on brlA by examining mRNA levels of brlA in OEvelB strains grown in liquid submerged culture, where neither WT nor OEvelB strains developed conidiophores (not shown). As shown in Fig. 2C, OEvelB was not sufficient to induce the activation of brlA, indicating that VelB is necessary, and conditionally sufficient for activating conidiation in A. nidulans.

VelB Negatively Controls Germination of Conidia
To test the role of velB in spore germination, we examined germination rates of conidia of WT, DvelB and complemented strains in the presence of 1% glucose. At 4 h of incubation, germination rates of WT, DvelB and complemented strains were about 0%, 60% and 0%, respectively. All the DvelB conidia formed germ tubes by 7 h, whereas all conidia of WT and complemented strains germinated by 8 h (Fig. 3A). In further tests, conidia of WT, DvelB and complemented strains were inoculated onto solid media without an external carbon source. At 5 h after inoculation, some of the DvelB mutant conidia began to form germ tubes but nano-LC-MS system. Polypeptides identified from the bands of the VelB-FLAG affinity purification from conidia (A), ascospores (B) and hyphae (C) are shown. doi:10.1371/journal.pone.0045935.g005 these were not detectable in WT and complemented strains (Fig. 3B). These results indicate that VelB is negatively associated with conidial germination.
Previous studies suggested that GanB(a)-mediated signal transduction plays a crucial role in controlling germination and pigmentation via cAMP/PKA signaling pathway, and is attenuated by the regulator of G-protein signaling (RGS) protein RgsA [24,[37][38][39]. Similar to DrgsA, a dominant activating GanB mutant allele caused enhanced accumulation of brown pigment(s) and increased conidial germination rates in the absence of a carbon source [24,37]. To study a potential relationship between VelB and the GanB signaling pathway, we generated the DvelB DganB double mutant and compared its phenotypes including brown pigment production and conidial germination with the DvelB mutant. The DvelB DganB double mutant showed reduced accumulation of brown pigment compared to the DvelB single mutant. Similar to DvelB, however, the DvelB DganB double mutant exhibited reduced radial growth. As shown in Fig. 3C, although no WT and DganB conidia germinated, about 60% of the DvelB DganB double mutant conidia could germinate at 4 h of incubation. Taken together, these results suggest that, VelB may play an independent role to GanB in brown-pigment(s) production, but function downstream of GanB-mediated signaling in regulating conidial germination (Fig. 3D).

Inter-dependence of VelB and VosA
Previous studies showed that both VelB and VosA are needed for sporogenesis and trehalose biogenesis in A. nidulans [16]. To examine the potential inter-dependent relationship between VelB and VosA, we generated the vosA and velB double deletion mutant and checked the spore viability, trehalose amount and mRNA levels of tpsA in WT and mutants. When point inoculated on solid medium (Fig. 4A), the DvelB DvosA double mutant showed reduced conidiation compared to the WT and DvosA single mutant, but increased the number of conidia compared to the DvelB single mutant. These results suggest that VelB (activator) and VosA (repressor) provide independent inputs to the final levels of asexual development and the double deletion results in an inter-mediate phenotype (Fig. 4B). The conidia of both single and double mutants displayed severe viability defects (Fig. 4C). We measured trehalose amounts in 2 day old fresh conidia and found that trehalose was undetectable in the double deletion mutant (Fig. 4D). Northern blot of tpsA, which is associated with the synthesis of trehalose, revealed that both single and double deletion mutants exhibited reduced tpsA transcript levels in conidia (Fig. 4E). We also examined the tolerance of the spores to oxidative and UV stresses and found that the DvelB DvosA mutant conidia, similar to the DvelB or DvosA single mutant conidia, were more sensitive to oxidative and UV stresses compared to WT conidia (not shown). Moreover, we found that the absence of vosA, similar to DvelB, caused increased conidial germination rates in the presence or absence of carbon source (not shown). We also observed that germination rates of the DvelB DvosA mutant conidia were similar to the DvelB or DvosA single mutant conidia regardless of the presence or absence of an external carbon source. Taken together, these results suggest that VelB and VosA play an inter-dependent (not additive) role in sporogenesis, trehalose biogenesis and controlling conidial germination.

VelB Primarily Interacts with VosA in Asexual and Sexual Spores
VelB has been shown to form not only the VelB-VeA-LaeA heterotrimeric complex, which coordinates secondary metabolism and development, but also the VelB-VosA heterodimer, which may be the functional unit of trehalose biosynthesis and spore maturation [15,16]. The C-terminal region of VelB interacts with VeA or VosA [15]. Genetic data for the inter-dependent role of VelB and VosA in spores, and high accumulation of both vosA and velB transcripts in asexual and sexual spores [15], led us to hypothesize that VelB mainly interacts with VosA and forms the functional hetero-dimeric complex in the spores. To test this, we first generated the strains expressing the VelB protein fused with a 3xFLAG C-terminal tag in the DvelB background in A. nidulans and examined the levels of the VelB protein throughout the life cycle. Through Western blot with FLAG antibody, we found that the VelB::3XFLAG protein is clearly detectable during vegetative growth and early asexual development, and accumulates at high levels in conidia (Fig. S1). We confirmed that the VosA protein highly accumulates in conidia, too (Fig. S1). We then collected conidia and ascospores from 2 and 7 day, respectively, grown colonies of the strain expressing a functional VelB::3xFLAG fusion driven by the velB native promoter and identified the VelB interacting protein(s). We found that the VelB::3XFLAG immunopull-down predominantly isolated VosA and VelB, but not VeA or LaeA, in both asexual and sexual spores (Fig. 5 A&B). We further examined the VelB interacting protein(s) in hyphae and found that VelB::3XFLAG isolates VeA, but not VosA (Fig. 5C). These results suggest that VelB interacts with VeA (hyphae) or VosA (spores), and plays a differential role in controlling various biological processes.

AbaA Regulates velB and vosA Expression
During asexual development, mRNA levels of velB and vosA increase during the formation and maturation of conidia [13,15]. Through sequence analysis, we found that three and one AbaA response elements (AREs, 59-CATTCY-39) are located at the promoter and at N-terminal ORF regions, respectively, of velB, and hypothesized that AbaA directly binds and regulates velB (and vosA) expression in the late asexual development phases. To test this hypothesis, levels of velB mRNA were examined in various mutants carrying the brlA42, abaA14, wetA6. In the brlA42 and abaA14 mutants, velB mRNA accumulation decreased compared to WT, but in the wetA6 mutant, levels of velB mRNA were comparable to WT, suggesting that brlA and/or abaA are required for velB mRNA accumulation (Fig. 6A). To further test a potential direct activating role of BrlA and/or AbaA in velB expression, we examined levels of velB mRNA in the alcA(p)::brlA, alcA(p)::brlA abaA14 and alcA(p)::abaA mutants. As shown in Fig. 6B, velB mRNA highly accumulated at 6 h upon abaA overexpression. Levels of velB mRNA slightly increased at 24 h after brlA overexpression, which was dependent on the WT abaA allele, as velB mRNA was not induced in the alcA(p)::brlA abaA14 mutant. These results indicate that AbaA, not BrlA, is necessary and sufficient for velB mRNA expression.
We then investigated whether AbaA directly binds to the velB and vosA promoter regions by EMSA. We designed and used two probes covering AREs in the velB promoter and one probe covering ARE in the vosA promoter (Fig. 6C). As shown in Fig. 6D, the GST-AabA fusion protein binds to all three probes containing ARE whereas the GST protein alone failed to bind to these probes. To further corroborate these AbaA-DNA interactions, we designed another probe with the mutated ARE motif. The GST-AbaA fusion protein could not bind to these mutated probes (data not shown). Moreover, only the unlabeled WT probes, but not mutated probes, competed during the complex formation with the labeled WT ARE probes (Fig. 6C). These results demonstrate that AbaA directly binds to the promoter regions of velB/vosA and regulates velB/vosA mRNA expression.

Discussion
The velvet proteins are fungi-specific multifunctional regulators that control development and secondary metabolism in various filamentous and dimorphic fungi [14,40]. In A. nidulans, VelB plays an essential role in sexual development, the mycotoxin sterigmatocystin production and trehalose biosynthesis [15,16]. VelB homologues have been characterized in many fungi and are reported to regulate: sporulation and secondary metabolite production in Fusarium fujikuroi [41]; production of the trichothecene and zearalenone mycotoxins, pathogenicity, and sexual reproduction in Fusarium graminearum [42]; cell morphology and spore formation in Histoplasma capsulatum [43]; trehalose biosynthesis and conidiation in Aspergillus fumigatus (Park and Yu, unpublished); and production of the carcinogenic mycotoxin aflatoxin and formation of sclerotia in Aspergillus flavus (Park, Lovendahl and Yu, unpublished).
Our characterization of VelB in asexual development and sporogenesis indicates that VelB acts as a positive regulator of asexual development, and an essential component for trehalose biogenesis in spores. Previous studies suggested that the VosA-VelB heterodimer represses asexual development during vegetative growth in liquid culture and early asexual development in the dark [16]. Unlike DvosA strains, however, the DvelB mutant (i) cannot produce conidiophores in liquid submerged culture; (ii) produces a lower number of asexual spores than WT; and (iii) exhibits lowered brlA mRNA accumulation in early phase of asexual development. In addition, unlike the DvosA single mutant, the DvelB DvosA double mutant does not develop conidiophores in liquid submerged culture, i.e., VelB is epistatic to VosA. These results suggest that VelB is required for the hyper-conidiation in the absence of vosA, further supporting the proposed positive role of VelB in asexual development. Taken together, we propose that the VelB monomer or homo-dimer functions as a positive regulator of asexual development, whereas the VosA homo-dimer plays a negative regulatory role in conidiation during vegetative growth and the early phase of conidiophore formation (see Fig. 7).
VelB participates in the formation of multiple velvet complexes including VelB-VeA and VelB-VeA-LaeA that play differential roles in sexual development and secondary metabolism [14,15]. Our studies suggest that in hyphae VelB predominantly interacts with VeA and may form a VeA-VelB complex (Fig. 5C). As shown in Figs 1&2, however, VelB's function in asexual spore formation does not require the physical interaction with VeA, as VeA1 cannot physically interact with VelB (Bayram et al., 2008). These data suggest that VelB and VeA may play a redundant role in controlling asexual spore formation, but an inter-dependent role in sexual development by forming a VelB-VeA functional unit [15]. Similar to veA, moreover, the deletion of velB resulted in reduced colony size and elevated accumulation of brown pigment(s). The phenotypes, accumulation of brown pigment(s) and reduced colony growth, of the DveA or DvelB single mutant were similar to those of the DveA DvelB double mutant, suggesting that VeA and VelB also function inter-dependently to regulate brown pigmentation and colony growth. In addition, VeA and VelB control secondary metabolism by forming trimeric complexes with the master regulator of secondary metabolism LaeA [15].
It appears that, in addition to activating conidiation, VelB is required for proper control of conidial germination. Trehalose is not only required for the acquisition of tolerance to various stresses but also utilized in glycolysis, spore maturation and germination [11,44]. Trehalose is rapidly degraded during conidial germination and the absence of the trehalose-6-phosphate synthase gene caused delayed conidial germination in aspergilli [11,[45][46][47]. However, exogenous trehalose supplemented in the medium had no effects on rescuing the delayed conidial germination caused by the deletion of trehalose-6-phosphate synthase in A. fumigatus, suggesting that trehalose-6-phosphate synthase genes, but not trehalose, may play an additional role in conidial germination [47]. In our present study, despite the lack of trehalose in the velB deletion mutant conidia, germination rates of DvelB conidia were increased (Fig. 3). These suggest the possibility of VelB playing a more complex role in governing spore germination, which may be associated the GanB(a)-mediated signal transduction. We propose that VelB functions as a negative regulator of conidial germination acting downstream of GanB-mediated signaling (Fig. 3D).
Previous studies demonstrated that the absence of velB or vosA resulted in the lack of trehalose in conidia and shortened viability of spores, and that the VosA-VelB complex is formed in the nucleus. These suggest that the VelB-VosA complex is the functional unit governing conidial maturation [13,16]. Results of our double mutant analyses are generally in agreement with the proposed inter-dependent role of VelB and VosA in spore viability, trehalose biosynthesis and conidial germination. Our data further demonstrate that VelB predominantly interacts with VosA in the asexual and sexual spores (Fig. 5). Taken together, we confirmed that the VelB-VosA complex is a major unit playing a vital role in controlling spore maturation (trehalose biogenesis and cell wall completion), long-term viability, dormancy and germination.
BrlA is mainly localized in vesicles metulae, and phialides, and young spores and regulates conidiogenesis during the early phase of conidiation [48,49]. However, brlA mRNA has not been detected in mature spores [50]. To complete sporogenesis turning off the expression of brlA is crucial for fungal spores. Uncontrolled production of brlA may inhibit growth of A. nidulans, likely causing a generalized metabolic shutdown, leading to an inability of cells to acquire nutrients from the growth medium. [51]. Previous study demonstrated that VosA exerts negative feedback regulation of conidiation, i.e., turns off brlA expression after spore formation, and the deletion of vosA resulted in high level accumulation of brlA mRNA in conidia [13]. Our preliminary genome wide analyses of in vivo VosA-DNA interactions indicate that the promoter regions of brlA are specifically enriched, suggesting that VosA may directly bind to the promoter of brlA and repress brlA expression in phialides and/or conidia.
Based on these data, we propose a genetic model depicting the roles and regulation of VelB and VosA during conidiogenesis. VelB is a multifunctional developmental regulator playing a pivotal role throughout the life cycle of A. nidulans. The interaction between VelB and its partner proteins (VelB, VeA or VosA) is time and/or cell-type specific and the multiple VelB-related complexes play differential roles in controlling fungal development and sporogenesis (see model in Fig. 7). During conidiogenessis, both velB and vosA are activated by AbaA in the vesicle and phialide [13]. High level production of VelB and VosA in developing cells would induce the formation of the VosA-VosA and VelB-VosA complexes. Genetic data in conjunction with VosA's localization in metulae, phialides and conidia suggest that the VosA monomer or homo-dimer turns off brlA expression in conidia [13]. The VelB-VosA hetero-dimer activates focal trehalose biogenesis in spores, controls spore wall integrity, and negatively regulates precocious conidial germination (Fig. 7). Further in-depth analyses of the roles of the individual velvet proteins and complexes, and the mechanisms of velvet-mediated regulation are in progress and will provide new insights into fungal development and secondary metabolism.