The veA gene is a light-dependent regulator governing development and secondary metabolism in Aspergillus nidulans. We have identified a putative bipartite nuclear localization signal (NLS) motif in the A. nidulans VeA amino acid sequence and demonstrated its functionality when expressed in yeast. Furthermore, migration of VeA to the nucleus was dependent on the importin α. This bipartite NLS is also functional when VeA is expressed in A. nidulans. Interestingly, we found that VeA migration to the nucleus is light-dependent. While in the dark VeA is located mainly in the nuclei, under light VeA is found abundantly in the cytoplasm. The VeA1 mutant protein (lacking the first 36 amino acids at the N-terminus) was found predominantly in the cytoplasm independent of illumination. This indicates that the truncated bipartite NLS in VeA1 is not functional and fails to respond to light. These results might explain the lack of the morphological light-dependent response in strains carrying the veA1 allele. We also evaluated the effect of light on production of the mycotoxin sterigmatocystin in a veA wild-type and the veA1 mutant strains and found that the highest amount of toxin was produced by the veA+ strain growing in the dark, condition favouring accumulation of VeA in the nucleus.
Most Aspergillus species disseminate efficiently by generating asexual spores called conidia. Other species, including the model filamentous fungus Aspergillus nidulans, also produce sexual spores called ascospores during their life cycle. Conidia are formed on specialized structures called conidiophores while ascospores are formed in sexual fruiting bodies denominated cleistothecia. Conidiophore formation is initiated from a thick-walled foot cell forming a stalk. The tip of the stalk forms a vesicle where elongated cells called sterigmata develop. From the sterigmata, long chains of airborne conidia are produced. During A. nidulans sexual stage, vegetative hyphae coil and fuse forming the cleistothecium. Large thick-walled cells called Hülle cells nurse the cleistothecial primordia contributing to the formation of the cleistothecium wall. Reproductive ascogenous hyphae proliferate generating ascospore-containing asci within the cleistothecia (Alexopoulos, 1962; Yager, 1992; Alexopoulos et al., 1996).
Aspergillus and other fungi are able to sense extracellular factors that are integrated by regulatory networks influencing a developmental adaptive response. This response balances vegetative growth and both asexual and sexual development (Adams and Yu, 1998; Calvo et al., 2001; 2002; Braus et al., 2002; Tsitsigiannis et al., 2004a). In A. nidulans, environmental factors such as light, pH, oxidative metabolism and nutritional factors direct morphogenesis towards asexual or sexual development (Clutterbuck, 1977; Mooney and Yager, 1990; Mooney et al., 1990; Skromne et al., 1995; Adams and Yu, 1998; Peñalva and Arst, 2004). Among these environmental factors, light has a major effect on A. nidulans morphological development triggering asexual differentiation, whereas in the absence of light the fungus differentiates forming sexual fruiting bodies (Yager, 1992; Champe et al., 1994). In this fungus the velvet gene, veA, is a global regulator that mediates this developmental response to light (Yager, 1992). The veA deletion in A. nidulans blocks the formation of cleistothecia (Kim et al., 2002) while conidial production is increased (Kim et al., 2002; Kato et al., 2003). In addition, strains carrying a veA1 mutant allele have been, and are, extensively used in the Aspergillus research. The veA1 allele has a point mutation in the first ATG and consequently the first methionine in the VeA1 protein corresponds to M37 of the wild type. The veA1 genetic background causes loss of VeA light-dependent response, resulting in increased conidiation under light and dark growth conditions, and presenting a delayed and reduced sexual development. This fact suggests that the N-terminal region of the VeA protein is vital for the A. nidulans light-regulated morphogenesis. The VeA N-terminal region contains two putative nuclear localization signal (NLS) motifs. One of them, a predicted pat7 motif, was indicated by Kim et al. (2002) from amino acid K41 to C46 of the deduced amino acid VeA sequence. The functionality of this putative pat7 motif has not been investigated. In this report we present a second NSL motif, a predicted bipartite NLS that extends from K28 to R44 amino acids (partially overlapping pat7). We were particularly interested in this putative bipartite NLS as, based on the information described above, the ‘blind-to-light’ VeA1 protein would present a truncation of this bipartite NLS motif.
In addition to the role of veA in morphogenesis, we have previously described veA to be essential for normal production of mycotoxin and other secondary metabolites (Kato et al., 2003; Calvo et al., 2004; Duran et al., 2006). The possible connection between VeA subcellular localization and mycotoxin production has not been investigated until now.
Here we report a novel study on the subcellular localization of VeA and its response to light in A. nidulans. We found a similar bipartite NLS conserved in other Aspergillus spp., including the human pathogen Aspergillus fumigatus and the aflatoxin producers Aspergillus parasiticus and A. flavus. Therefore, the implications of this study in the model fungus could extend to other fungal species, contributing to a better understanding of the light-sensing mechanisms as well as of the mechanisms that are part of the adaptive response to this stimulus.
Aspergillus nidulans VeA has a putative classical bipartite NLS motif also present in other Aspergillus spp.
We found in VeA a conserved bipartite NLS-like motif, related to that described at the SV40 large tumour antigen or nucleoplasmine NLSs. Although similar, the VeA NLS does not completely fit in with the consensus established by Fontes et al. (2000; 2003a) for bipartite NLSs. This putative NLS is located at the N-terminal of the VeA-deduced amino acid sequence. We analysed for the presence of VeA and conservation of this putative bipartite NLS in Aspergillus spp. and across fungal genera (Fig. 1). We found that VeA was not found in strictly yeast organisms such as Saccharomyces cerevisiae or Schizosaccharomyces pombe. However, VeA was found in filamentous fungi other than Aspergilli, for example in Fusarium graminerum, Magnaporthe grisea or the dimorphic fungus Histoplasma capsulatum (K. Myung, M.R. Duval and A.M. Calvo, unpublished). A similar bipartite NLS was found in all the Aspergillus VeA sequences analysed (Fig. 1): A. flavus, A. oryzae, A. parasiticus, A. terreus, A. fumigatus, A. clavatus along with A. nidulans. This bipartite NLS seems exclusive of Aspergillus spp. as when VeA from other fungal genera were analysed the bipartite motif was not found (with the exception of H. capsulatum, Neosartorya fischeri, Uncinocarpus reesii and Coccidioides immitis).
Aspergillus nidulans VeA NLS is functional
To test whether this NLS is functional we first generated a series of constructs containing different alleles of A. nidulans veA (veA, ΔNLSveA and veA1) fused to the gene encoding the green fluorescent protein (gfp) as indicated in the diagram in Fig. 2A. Additional constructs with the monopartite SV40 large T antigen NLS to these fusion proteins were included as internal controls for nuclear localization. The contructs were introduced in S. cerevisiae (S. cerevisiae does not have an endogenous VeA, see previous section) using centromeric plasmids, obtaining strains expressing: GFP::VeA; NLSsv40::GFP::VeA; GFP::ΔNLS-VeA; NLSsv40::GFP::ΔNLS-VeA; GFP::VeA1; and NLSsv40::GFP::VeA1. Figure 2B shows that VeA was found accumulating mainly in the nuclei, while the ΔNLS-VeA was distributed equally between nucleus and cytoplasm, as occurs for GFP single constructs (not shown). These results indicate that VeA has a functional NLS at the N-terminus and is recognized by the yeast nuclear import machinery. Interestingly, the GFP::VeA1 chimera showed cytoplasmic localization (Fig. 2C), contrasting with the positive control NLSsv40::GFP::VeA1 that located to the nucleus, as expected (under identical experimental conditions). These results strongly indicate that the first 36 amino acids missing in VeA1 mutant protein are crucial for VeA localization to the nucleus. Therefore, the truncated bipartite NLS and the pat7 putative NLS present in VeA1 (described by Kim et al., 2002) are not sufficient for effective transport to the nucleus of yeast cells. This result is relevant as most laboratories using A. nidulans as a model organism utilize the veA1 mutation to induce conidiation in the dark, despite the lack of knowledge over molecular mechanism(s) altered in the veA1 mutant background.
VeA migration to the nucleus is dependent on the importin α carrier protein Srp1p/KapA
The data shown above support the hypothesis that VeA contains a classical bipartite NLS that could be specifically recognized by the general nuclear transporter importin α. The importin α homologue to Srp1p was found in A. nidulans and denominated KapA (AF465210; J. Fernandez-Martinez and E.A. Espeso, unpublished; Fernandez-Martinez et al., 2003). There are no importin α mutants currently available in A. nidulans. We used then the S. cerevisiae thermosensitive srp1-31 importin α mutant allele to investigate whether VeA transport to the nucleus was dependent on this protein carrier. Our results showed that the preferential VeA nuclear localization was indeed dependent on the activity of the importin α (Fig. 3A). At the permissive temperature (25°C) for the mutated form srp1-31p VeA was efficiently transported to the nucleus, while at restrictive temperature (37°C), at which srp1-31p is inactive, VeA nuclear localization was lost. Expression of GFP::VeA in wild-type yeast cells resulted in VeA nuclear localization also at 37°C.
We analysed the importin α recognition of A. nidulans VeA NLS by using the two-hybrid system (Fig. 3B). Several VeA forms were analysed as baits: 1–573 (wild-type) and truncated forms 45–573 (ΔNLS-VeA), 186–573 and 37–573 (VeA1). Our results indicated that the 1–573 wild-type efficiently interacts with KapA, while ΔNLS-VeA and VeA1 showed a weak interaction with KapA. The way in which the wild-type VeA form interacts with KapA and the fact that ΔNLS-VeA and VeA1 presented only a weak interaction is in agreement with our microscopic observation described above and evidences the bipartite nature of the NLS, as VeA1 lacks half of the signal. We then analysed in vitro the previously detected interaction between N-terminus region of velvet and the importin α. In agreement with the preceding results, KapA50 (the truncated version lacking the auto-inhibitory domain) interacted with the wild-type VeA protein and showed a slightly weaker interaction with proteins having the region between residues 37 and 186 (Fig. 3C, and confirmation by densitometry, data not shown). These results suggest that the region comprising residues 37–186 might participate in the KapA nuclear-mediated transport of velvet and that in vivo the bipartite NLS might be post-transcriptionally modified to improve binding to the nuclear carrier.
No classical consensus export signals (NES) were found in the VeA sequence. Furthermore, we tested the possible interaction of VeA with the exportin 1 KapK (Accession No. AY555733, Bernreiter et al., 2006), the A. nidulans CRM1p orthologue (Kutay and Güttinger, 2005), by the two-hybrid assay. The results indicated no interaction between these two proteins (data not shown).
VeA is also found in the nucleus when expressed in A. nidulans
In order to express VeA fused to GFP in A. nidulans, we generated a veA::gfp construct by a three-way PCR-based 3′ tagging protocol (described in Experimental procedures). This construct was inserted in the A. nidulans veA locus as illustrated in Fig. 4A, where veA expression is governed by its own promoter. The correct integrations were confirmed by PCR and Southern blot analysis (not shown). The cultures were incubated overnight in the dark. In Fig. 4B we show that VeA is found into the nuclei in A. nidulans as indicated by the nucleus-specific 4′,6-diamidino-2-phenylindole (DAPI) staining.
A similar strategy was used to express the VeA1::GFP fusion (Fig. 4C). In contrast to the wild-type VeA, the truncated VeA1 protein was found in the cytoplasm (Fig. 4D), associated in part with elongated structures. When treated with benomyl the elongated structures remained (data not shown), suggesting that these structures are not part of the cell tubulin cytoskeleton. Further investigation is needed to elucidate the nature of these structures and whether they are associated with the light-sensing apparatus in A. nidulans.
Additionally we performed Western analysis using antibodies against GFP to visualize the VeA–GFP-tagged versions, detecting a 90 kDa band in both veA+ and veA1 backgrounds as a result of the fusion between velvet and GFP evidencing that the VeA1::GFP truncated allele was being expressed in Aspergillus (Fig. S1). Wild-type VeA was previously detected by Western blot by Toews et al. (2004).
All the transformants expressing the VeA::GFP chimera showed wild-type phenotype in their light-dependent morphological development (Fig. S2) and also regarding the production of the mycotoxin sterigmatocystin (ST) (Fig. 5), indicating that the VeA::GFP fusion protein is functional. Although ST production still occurs in a veA1 mutant background (i.e. Hicks et al., 1997; Shimizu and Keller, 2001), the absence of veA results in a complete blockage in ST biosynthesis (Kato et al., 2003). Using our velvet–GFP fused strains, we detected that the ST amount produced by the veA1 mutant strain in the dark was similar to that of the veA+ wild-type strain under light growth conditions, and notably lower than the veA+ in the cultures grown in the absence of light. These results led us to examine whether the differential response to light in both morphogenesis and secondary metabolism was a consequence of changes in VeA subcellular localization depending on this environmental factor.
VeA subcellular localization is light-dependent
veA is a light-dependent regulatory gene governing development and mycotoxin production in A. nidulans (Yager, 1992; Kim et al., 2002; Kato et al., 2003). Interestingly, our results showed the light-dependent subcellular localization of VeA (Fig. 6A and B). While in the dark VeA is located mainly into the nuclei, when grown in the presence of light VeA distributes between nucleoplasm and cytoplasm. Noticeably, the cytoplasmic VeA::GFP was also found associated with filamentous bodies, coinciding with areas highlighted by DAPI staining (Fig. 6C). A red-light phytochrome (FphA) is also associated to large particles in the cytoplasm in A. nidulans (Blumenstein et al., 2005). In plants the phytochrome is associated with electrodense particles (McCurdy and Pratt, 1986). It is possible that VeA could be forming part of large protein complexes in the cytoplasm with the phytochrome in the light and associated with cell organelles.
Quantification of the fluorescence intensity (Wasabi Version 1.4, Hammatsu Phototonics) supported our observations (Fig. 6B). In the dark, the nuclear/cytoplasmic fluorescence ratio was approximately 1.6, while in the light the ratio was approximately 1.1 (samples n = 10). Independent experiments were repeated three times obtaining similar results. The results from the quantification of fluorescence in the VeA1::GFP cultures were similar to those in VeA::GFP light cultures (data not shown).
In our study, VeA1 showed cytoplasmic localization independently of light, indicating that the absence of the first 36 amino acids in the VeA1 form causes a failure to respond to the light stimulus.
Differently from A. nidulans, in our experiments in yeast the nuclear localization of VeA was light-independent (data not shown). One possibility is that VeA is regulated negatively in the light in A. nidulans by an unknown molecule or complex, and that this molecule or complex has been lost in yeast, resulting in VeA nuclear localization in either light or dark.
Accumulation of VeA in the nucleus varies with changes in illumination regimen
In order to further understand the VeA subcellular localization in response to the presence or absence of light, we carried out shift experiments from light to dark and from dark to light. Our results showed that the intensity of the nuclear green fluorescence varied when the cultures were exposed to changes in illumination (Fig. 7), increasing when the culture was shifted from light to dark and decreasing when the culture was shifted from dark to light (the fluorescence intensity in the cytoplasm did not increase; data not shown). These results indicate that the VeA protein tends to accumulate in the nucleus when the fungus no longer has a light source. When dark cultures were exposed to light, the VeA::GFP protein nuclear levels decreased. A possible explanation to these observations is that VeA might be subjected to degradation in the presence of light.
Effect of blue light and red light on VeA subcellular localization
We tested the possible effects of blue light and red light on VeA subcellular localization. Our experiments showed that blue light had a similar effect to that observed with white light, preventing an efficient accumulation of VeA::GFP in the nuclei (Fig. 8). The effect of the exposure to red light was much moderated, still allowing protein accumulation in the nuclei, similarly to that observed in dark cultures (Fig. 8).
In A. nidulans the veA gene regulates a balance between sexual and asexual development in response to light. In the light the fungus develops forming conidiophores that bare airborne conidia while in the dark morphogenesis is directed to the formation of more resistant sexual fruiting bodies called cleistothecia. In spite of its importance, until now no previous reports have shown evidence of how VeA regulates development and how VeA integrates the light response.
We identified a putative bipartite NLS motif in the A. nidulans VeA-deduced amino acid sequence and demonstrated that this NLS is functional, indicating that VeA might exert its function into the nuclei, through interactions with other elements, protein and/or DNA. Furthermore, we have shown that the migration of VeA to the nucleus is dependent on the importin α, KapA, a carrier protein homologue to the S. cerevisiae Srp1p importin α protein, and demonstrated that this bipartite NLS, subject of our study, is necessary for normal interaction between VeA and KapA. In addition, our two-hybrid and in vitro results suggest the possibility of a post-transcriptional modification of VeA in vivo that could condition the interaction with KapA. Phosphorylation mechanisms have been described to influence the binding of NLS with the importin α (Fontes et al., 2003b). We have found several consensus phosphorylation sites in the VeA amino acid sequence (A.M. Calvo, unpublished) and we are currently investigating whether VeA is phosphorylated in A. nidulans. In our study no evidence of VeA nuclear export was observed .
Not only did we find VeA conserved in other fungal species, even across fungal genera, we also found a similar bipartite NLS motif conserved among Aspergillus species, including important plant and human pathogens (Fig. 1), and industrially important fungi, suggesting that VeA could be part of an Aspergillus-specific light-response mechanism. The closely related U. reesii and pathogenic N. fischeri, H. capsulatum and C. immitis were an exception presenting a VeA protein with a similar NLS motif. The functionality of these NLSs in other Aspergillus species remains to be determined. NLSs have been experimentally determined for fewer than 10% of known nuclear proteins (Nair et al., 2003) and therefore it is possible that these NLSs could also be functional. Interestingly, VeA was not found in fungi with a strictly yeast form.
It is also important to note that the VeA1-truncated protein form present in many strains used in the Aspergillus research community lacking part of the NLS shows reduced interaction with the importin α, and consequently, a deficient transport to the nucleus compared with the VeA wild-type protein. This finding strongly suggests that differences in VeA1 subcellular localization could account for the striking alteration of sexual/asexual ratio in veA1 strain as well as metabolic differences compared with the veA wild-type strain.
Our results indicate that the subcellular localization of VeA depends on the presence or the absence of light. In the dark most of the VeA protein was found in the nuclei of fungal cells, while in the presence of light the level of VeA accumulation in the nucleus was lower compared with that observed in the dark. We also found this trend when the cultures were shifted from light to dark and from dark to light. Light cultures that were shifted to dark experienced an increase of VeA in the nuclei, while dark cultures exposed to light showed a reduction of this protein in the nucleus, without evidence of VeA nuclear export. Based on these observations it is possible that VeA protein degradation could be stimulated in the presence of light. This will possibly be the subject of future studies.
In a recent study a phytochrome denominated FphA that acts as a red-light sensor has been reported in A. nidulans (Blumenstein et al., 2005). According to previous work, red light represses sexual development and induces conidiation in A. nidulans (Mooney and Yager, 1990; Blumenstein et al., 2005; Idnurm and Heitman, 2005). The FphA protein was found exclusively in the cytoplasm, evidence that suggests that at least part of the red-light photoreception occurs in the cytoplasm (Blumenstein et al., 2005). Although our study showed that red light had a mild effect on the VeA subcellular localization, it is possible that VeA could interact with FphA. The fact that the fphA mutant presents an increase in cleistothecial formation in the light with respect to the wild type (Blumenstein et al., 2005) supports our hypothesis. Another line of evidence supporting our hypothesis is the fact that the fphA mutant phenotype is only observed in VeA strains and not in VeA1 (Blumenstein et al., 2005). According to our results, the cytoplasmic localization of the VeA1 protein form indicates that the truncated bipartite NLS in VeA1 is not functional and fails to efficiently accumulate in the nucleus independently of illumination and most likely independently of the phytochrome FphA. We are currently studying the relationship between FphA and VeA. It is interesting that both veA and fphA are conserved in other filamentous fungi but are not found in yeast genomes suggesting that this light-response mechanism is exclusive to filamentous fungi. In our study, the nuclear localization of VeA when expressed in S. cerevisiae was light-independent.
Additional elements could participate in a complex light response. Although the fphA mutant presented an increase in cleistothecial formation under the light, the number of cleistothecia was still considerablely lower than that found in the wild type growing in the dark, suggesting additional sensor systems (Blumenstein et al., 2005). The fluG gene, first described by Lee and Adams (1994), seems to be one of those candidate elements: Yager et al. (1998) described fluG mutants with restored light-dependent conidiation in the veA1 (which conidiates independently in red light), indicating a possible involvement of the FluG protein in the red light response. The functional dependence of FluG on FphA is unknown. Furthermore, the data presented by Blumenstein et al. (2005) also suggested additional sensors for red light besides FphA.
The response to blue light has been studied in detail in other fungi, including the filamentous fungus Neurospora crassa (Froehlich et al., 2002; He et al., 2002; Liu et al., 2003). Response to blue light has also been indicated in A. nidulans (Yager et al., 1998), reporting a bliA1 mutant that lost the blue-light developmental response. The WC-1 and WC-2 proteins are required in the blue light-sensing mechanism in N. crassa (Froehlich et al., 2002; He et al., 2002; Liu et al., 2003). Although orthologues of both white collar genes were found in the A. nidulans genome, much of the response to blue light remains unknown in this model fungus. In our experiment we found a strong effect of blue light in preventing efficient VeA accumulation to the nuclei, similar to that shown in the presence of white light. Further research is needed to elucidate the possible interactions between VeA and proteins responsive to blue light in A. nidulans.
In A. nidulans, veA is required not only for normal morphological development but also for the production of secondary metabolites. Our group discovered a connection between veA and secondary metabolism, among others, the production of the mycotoxin ST (Kato et al., 2003). The results here presented suggest that although a small amount of VeA in the nucleus (such of VeA1 and VeA in the light) could be sufficient to activate the ST gene cluster, the highest amount of toxin was produced by the veA+ strain growing in the dark, a condition that favoured the accumulation of VeA protein in the nucleus. In previous studies we demonstrated that VeA regulates the expression of the ST-specific transcription factor gene aflR as well as expression of enzymatic genes in the ST gene cluster (Kato et al., 2003) supporting a VeA regulatory action in the nucleus regarding mycotoxin production.
Accumulation of VeA in the nucleus in the dark must trigger a series of events that lead to the developmental stage observed in the A. nidulans veA wild-type strain. The VeA mechanism of action in the nucleus is at this time unclear, as VeA does not have homology with any known transcription factor or other proteins with known function. However, it has been demonstrated that VeA affects, directly or indirectly, the transcription of genes known to be involved in sexual and asexual development. In a previous report (Kato et al., 2003) we showed that veA affects the α/β transcript ratio of brlA, encoding a transcription factor that controls conidiation events in A. nidulans (Han et al., 1993; Han and Adams, 2001). Other developmental genes linked to veA are the Psi factor oxilipin genes that control a balance between asexual and sexual development (Calvo et al., 2001; Tsitsigiannis et al., 2004a,b; 2005). Deletion of veA completely prevents the expression of the oxylipin gene ppoA, coinciding with a blockage of sexual development and with an increase in conidiation (Tsitsigiannis et al., 2004a). Additionally, the triple null mutation ppoAppoBppoC alters veA expression, suggesting a regulatory loop between ppo genes and veA (Tsitsigiannis et al., 2005). The expression of the sexual development transcription factors nsdD and steA was only slightly altered in the veA mutant (Kato et al., 2003). However, it is possible that VeA regulation of these developmental factors could also occur post-transcriptionally.
The discovery of VeA light-dependent subcellular mobility is an important keystone to understand the mode of action through which Aspergilli exercise a complex developmental response to light stimulus thus adapting to their environment.
Saccharomyces cerevisiae procedures
The S. cerevisiae strains used in this study were W303-1a (MATa ura3-1 leu2-3112 his3-11 trp1-1 can1-100) and its isogenic derivative carrying the srp1-31 thermosensitive mutation (Fernandez-Martinez et al., 2003) for the microscopy analysis. The strain Y187 (Clontech, MATα ura3-52 his3-200 ade2-101 trp1-901 leu2-3112 gal4Δmet-gal80ΔURA3::GAL1UAS -GAL1TATA -lacZ MEL1) was used for the two-hybrid system study. Strains were cultured in dextrose minimum medium with the appropriate supplements for plasmid selection maintenance.
All the veA alleles were expressed as N-terminal sGFP fusion proteins under the control of the ADH1 promoter, in the centromeric plasmid pRS313 (Sikorski and Hieter, 1989). Plasmid construction for VeA::GFP expression in yeast was performed as follows. First cDNAs corresponding to the veA, veAΔNLS and veA1 alleles were amplified with the forward primers 79YGFPWTF (all the primers used in this study are listed in Table 1), 97YGFPNSL and 110YGFP, respectively, and the reverse primer 80YGFPRSmaI, and cloned into the PstI/SmaI sites of pBlueScript SK+ containing the ADH1 promoter [ADH1(p)] and into a similar plasmid with both ADH1(p) plus NLSSV40 [the fragments corresponding to ADH1(p) and ADH1(p)+ NLSSV40 were previously inserted into XhoI/EcoRI sites in pBlueScript SK+Fernandez-Martinez et al., 2003]. XhoI/SmaI fragments were released from the resulting plasmids and ligated into XhoI/EcoRV sites of a modified pRS323 containing the ADH1 terminator (as described in Fernandez-Martinez et al., 2003). Finally, a fragment corresponding to the gene encoding GFP was inserted into a unique BamHI site in the multiclonal site of each plasmid (a BamHI site in the cDNAs was previously eliminated by a silent mutation TCcGATCCC). The plasmids obtained are listed in Table 2. pAMCA.1GFP, PAMCA2GFP, pAMC43GFP, pAMC45GFP, pAMC52.I2GFP and pAMC53.NLS2GFP were used to transform S. cerevisiae W303-1a. pAMCA.1GFP and pAMCA.2GFP were also used to transform the S. cerevisiae srp1-31 strain.
Plasmid construction for the yeast two-hybrid system studies was performed as follows. The cDNAs corresponding to the veA, veAΔNLS and veA1 alleles were amplified with the forward primers 92YtwohybrWTF, 93TwoHybrΔNLS and 109YtwohybrveA1d, respectively, and the reverse primer 95YTwohybrR. The PCR products corresponding to the veA and veAΔNLS were digested with NcoI and PstI and inserted in NcoI/PstI sites of pGBKT7. The cDNA fragment corresponding to veA1 was digested with NdeI and PstI and inserted in NdeI/PstI sites of pGBKT7. pGBKT7 is a suitable partner for pACT2 (A. nidulans cDNA encoding for KapA residues 83–551 was previously introduced as a NcoI fragment in pACT2). The resulting plasmids are listed in Table 2. An N-terminus truncated veA allele, veA186-573, was generated by digesting pAMC46 (Table 2) with EcoRI (the EcoRI in veA is in frame with the EcoRI site of pGBKT7) and religation of the plasmid resulting in pAMC49. pAMC46, pAMC47, pAMC49 and pAMC54 were introduced in S. cerevisiae Y187. Yeast transformation and β-galactosidase assays were carried out as previously described (Vincent et al., 2002). For green fluorescence observation, yeast cells were cultured in SD medium until reaching a cell density of A600 = 0.6–0.8. Afterwards, cells were pelleted and resuspended in 0.05 of the original volume and mounted for visualization.
In vitro interaction analyses were performed using TNT (Promega) expressed VeA forms using the pGBKT7 derivatives for two-hybrid analysis, and bacterially expressed ZZ::KapA50 (83–551) and ZZ tag alone as control. In vitro synthesized VeA proteins we performed as described by the manufacturer (T7-TNT Promega) and labelled with Methionine-S35 (Amersham). Plasmid pQE80zz (a gift from Dr D. Görlich) was used to express ZZ::KapA50. The KapA50-coding fragment was PCR-amplified from cDNA using oligonucleotides α5 and α6 (Table 1) and cloned into the BamHI site of pQE80zz. Plasmids expressing either ZZ::KapA50 or ZZ were introduced into Escherichia coli DH1. Transformants were grown for 16 h at 30°C in LB broth plus 100 μg ml−1 ampicillin. Cultures were diluted 1/50 with fresh medium and further incubated at 30°C until cell density reached 0.6 units OD at 600 nm. Expression of ZZ::KapA50 or ZZ was induced by addition of 0.1 mM IPTG followed by 3 h incubation at 30°C. Cells were collected by centrifugation and resuspended in binding buffer (buffer B, 50 mM Tris HCl pH 7.5, 50 mM NaCl, 2 mM MgCl2) containing complete protease inhibitor cocktail (Roche Diagnostics). The cell suspension was lysed with a French Press. Samples were clarified by centrifugation for 15 min at 14 000 r.p.m., 4°C. Supernatants were stored at −80°C.
For the in vitro interaction analyses, IgG-sepharose (Roche) was loaded with either ZZ-tagged KapA50 or ZZ as follows: 20 μl of the matrix were incubated with 300 μg of an E. coli extract expressing ZZ or ZZ::KapA50 in HNA buffer (50 mM HEPES/KOH pH 7.5, 200 mM NaCl, 5 mM magnesium acetate and complete protease inhibitor cocktail), in a final volume of 1.5 ml. The matrix was washed three times with HNA buffer before utilization. For each experiment similar amounts of labelled S35-Methionine-VeA protein TNT mixtures were incubated for 16 h at 4°C in a rotary wheel with either ZZ or ZZ-KapA50; bound to the IgG-sepharose. Each experiment was washed three times with 1 ml each of HNA buffer. Protein samples were dissolved in SDS-PAGE loading buffer containing 1% v/v β-mercaptoethanol, boiled for 5 min and desolved on a 10% SDS-PAGE, stained with Coomasie blue, dried and exposed to S35-sensitive film.
Western blot analysis
Protein extracts were obtained from mycelia grown on GMM for 18 h in the dark. The mycelia were frozen on dry-ice and lyophilized. Proteins were extracted in the following buffer: 5 mM HEPES pH 7.5, 20 mM KCl, 1 mM EDTA pH 8.0, 10% NP40, 0.5 mM DTT, 1 mM Pefablock, 1 mM Pepstatin, 0.6 mM Leupeptin. Protein quantification was carried out by the commonly used Bradford method. Fifty micrograms of total protein from VeA+::GFP, VeA1::GFP and untagged VeA strains were used for Western analysis. The filter was probed with a mouse polyclonal primary antibody against GFP (Roche) in a 1:10.000 dilution; the secondary antibody was used in a 1:4.000 dilution (Jackson).
Insertion of the gfp tag in the veA locus in A. nidulans
We used the three-way PCR-based 3′ GFP tagging protocol described by Yang et al. (2004) to generate the veA::gfp transformation cassette. The method leads to an in-frame integration of GFP just before the stop codon of veA. The GFP cassette generated includes a hinge region encoding five Gly–Ala repeats in frame at the N-terminus of GFP, next to the A. fumigatus pyrG gene that was used as a selectable marker for fungal transformation. The use of A. fumigatus pyrG instead of the A. nidulans pyrG reduces integrations at the pyrG locus. In brief, as an initial step, three separate fragments were amplified. All the primers of this section are listed in Table 1. First, the GFP cassette was amplified by PCR, using primers GFP1 and GFP2 and using p1439 (containing GFP plus A. fumigatus pyrG) (the p1439 plasmid was kindly provided by Stephen Osmani) as a template. The second PCR fragment corresponding to the upstream targeting region (the 3′ end in veA-coding region) was PCR-amplified using primers GSP1 and GSP2 with pPK11 (containing the full-length veA-coding region) (gift from Larry Yager) as a template. The third amplified fragment contains the veA stop codon and the 3′ untranslated region and was amplified using primers GSP3 and GSP4 with FGSC4 genomic DNA as template. The overlapping ends of the three described PCR products allows a three-way PCR fusion, amplified by using all three fragments as a template in one PCR reaction where the primers GSP1 and GSP4 are used. All PCRs were performed using the Boehringer Manheim Expand Long Template PCR kit. In the case of veA1 strains, gfp tagging was carried out following the same procedure.
The final PCR product was used to transform two A. nidulans strains, one with a veA wild-type allele, RRMD3.4 (Table 3), and a second strain with a veA1 allele, FGSC A773 (Table 3). The fungal transformation was carried out as previously described (Miller et al., 1985). The integration of the tag occurs via a double-cross-over event. The complete list of A. nidulans strains used in this study is provided in Table 3.
Table 3. Aspergillus nidulans strains used in this study.
pyrG89; pyroA4; veA+
pyrG89; pyroA4; veA1
Fungal Genetics Stock Center
Aspergillus nidulans culture conditions
Aspergillus nidulans conidia from TRMD3.4.17 and TA773.B13 transformants and corresponding isogenic controls (Table 2) were inoculated onto coverslips and overlayed with minimal media (GMM, Calvo et al., 2002) plus the appropriate supplements (Käfer, 1977). The coverslips were incubated at 26°C and 37°C for 25 h (18 h and 20 h gave the same results) in the light (approximately 25 mE m−2 s−1) or in the dark. The experiment using blue (440–500 nm) and red (625–740 nm) light filters (Arbor Scientific) were carried out in the same manner as described above. For the shift experiment, the VeA::GFP strain was grown for 18 h in the light or in the dark and then shifted from light to dark and from dark to light. Samples were harvested and visualized at the shift time and 30 and 60 min after the shift.
All samples were prepared for microscopy analysis as described below. The experiment was repeated four times with similar results. Cultures for mycotoxin biosynthesis evaluation of A. nidulans veA wild-type strain versus veA1 mutant were performed as follows: conidia from RNKT5.1 and REE1 as well as from TRMD3.4.17 and TA773.B13 (Table 3) were inoculated on plates containing 25 ml of GMM agar. For each plate, a 5 ml top layer of cold but molten agar that contained 106 spores of the appropriate strain was added. For each strain, there was a minimum of three replicate plates. Strains were grown at 37°C in continuous light or dark. The experiment was performed in triplicate.
Phosphate-buffered saline (PBS) with 4% (v/v) paraformaldehyde was used to fix the cultures. After three washes in PBS, the coverslips were stained with DAPI (6 ng ml−1 in PBS) for 15 min followed by another three PBS washes. The subcellular distribution of green fluorescence was observed with an ORCA-ER digital camera (Hamamatsu) coupled to a NIKON E-600 microscope equipped with a 60× objective and 495 and 530 nm excitation and emission filters. No autofluorescence was observed. All images were taken using the same exposure and microscope settings.
Four cores (11 mm diameter) from each replicate were collected in a 50 ml Falcon tube, and ST was extracted by adding 5 ml of CHCl3 three consecutive times. The extracts were allowed to dry and then resuspended in 500 μl of CHCl3 before 15 μl of each extract was fractionated on a silica gel thin-layer chromatography (TLC) plate using a toluene-ethyl acetate-acetic acid [80:10:10 (v/v/v)] solvent system. The TLC plates were sprayed with aluminum chloride (15% in ethanol) to intensify ST fluorescence upon exposure to long-wave (365 nm) UV light and baked for 10 min at 80°C prior to being viewed. The approximate sensitivity of the assay was 25 ng. ST purchased from Sigma was used as a standard.
We thank Naoki Kato, Kyung Myung, Tracy Gold, Javier Fernández-Martínez and Elena Reoyo for their technical assistance. We are grateful to Stephen Osmani and Larry Yager for providing us with plasmids used in this study and to Masashi Nomura for the srp1-31 strain. We thank Northern Illinois University and the Ministerio de Educación y Ciencia (Grant BMC2003-00874 to E.A.E., L.A.-B. held a FPU) for providing funds to support this work.