The Aspergillus nidulans MAPK Module AnSte11-Ste50-Ste7-Fus3 Controls Development and Secondary Metabolism

The sexual Fus3 MAP kinase module of yeast is highly conserved in eukaryotes and transmits external signals from the plasma membrane to the nucleus. We show here that the module of the filamentous fungus Aspergillus nidulans (An) consists of the AnFus3 MAP kinase, the upstream kinases AnSte7 and AnSte11, and the AnSte50 adaptor. The fungal MAPK module controls the coordination of fungal development and secondary metabolite production. It lacks the membrane docking yeast Ste5 scaffold homolog; but, similar to yeast, the entire MAPK module's proteins interact with each other at the plasma membrane. AnFus3 is the only subunit with the potential to enter the nucleus from the nuclear envelope. AnFus3 interacts with the conserved nuclear transcription factor AnSte12 to initiate sexual development and phosphorylates VeA, which is a major regulatory protein required for sexual development and coordinated secondary metabolite production. Our data suggest that not only Fus3, but even the entire MAPK module complex of four physically interacting proteins, can migrate from plasma membrane to nuclear envelope.


Introduction
Eukaryotic organisms communicate between cell surface and nucleus to respond to environmental signals. The mitogenactivated protein kinase (MAPK) module consisting of a cascade of three protein kinases represents a highly conserved eukaryotic signal transduction system present from yeast to man. MAP3K phosphorylates a second kinase, MAP2K which itself phosphorylates the MAPK. This final kinase phosphorylates nuclear target proteins to activate appropriate gene expression [1,2].
The sexual pathway of the budding yeast Saccharomyces cerevisiae represents a paradigm for signal transduction in eukaryotes [3][4][5]. This MAP kinase pathway responds to pheromones and induces differentiation processes which trigger sexual mating of yeast [4,6]. The central complex of MAP3K Ste11, MAP2K Ste7 and MAPK Fus3 is assembled on the scaffold protein Ste5 as a hub to keep these kinases in a close proximity for enhanced relay of phosphorylation and thereby controls the flow of information [7]. Binding of pheromone to the transmembrane receptors Ste2 or Ste3, which are coupled to guanine nucleotide binding proteins (G protein, G protein coupled receptor: GPCR), initiates signal transduction. This induces the release of the Gbc subunit from the trimeric Gabc protein. The Ste5 RING domain binds to activated free Gbc complex and recruits the MAP kinase module Ste11-Ste7-Fus3 to the membrane [8][9][10] in close distance to the p21 activated kinase (PAK) Ste20. Preactivated Ste20 is localized in the membrane and initiates the kinase cascade system by phosphorylating the MAP3K Ste11 [4].
Ste50 represents a second adaptor which binds to the Opy2 membrane anchor and provides membrane association of the entire MAPK module. Ste50 mediated membrane localization is required for Ste11 activation [11,12]. The information is transmitted as phosphate signal from Ste11 via Ste7 to the MAPK Fus3. According to the current model phosphorylated Fus3 is released from the Ste5 scaffold complex and leaves the membrane associated complex [13][14][15]. Phosphorylated Fus3 crosses the cytoplasm and enters the nucleus where it phosphorylates target transcription factors as Ste12. Ste12 is necessary to activate the sexual pathway and also controls developmental processes [4,5].
Pheromone pathway genes have been studied in various fungi and are not only involved in sexual reproduction but also in fungal pathogenicity [16][17][18][19][20][21]. The Fus3 MAPK module is highly conserved in filamentous fungi with the exception that homologs for Ste5 are absent [22,23]. In the self-fertile model fungus Aspergillus nidulans, the Ste11 MAP3K homolog SteC (AnSte11) [24], the Fus3 MAPK homolog MpkB (AnFus3) [25], and the Ste12 homolog for the transcription factor SteA (AnSte12) [26] are necessary for sexual fruiting body formation, suggesting that there are similarities in the molecular function of the MAPK signal transduction as in yeast. A. nidulans grows vegetatively as a filament. When placed on a surface, after germination of the spores at least 12 hours of growth is required to establish developmental competence in response to external signals [27]. There are two developmental options: light supports the asexual and inhibits the sexual developmental pathway ( Figure 1A). AnFus3 is not only required for sexual development but also for the control of secondary metabolism which is a typical feature of many filamentous fungi [28]. Sexual development of A. nidulans is coordinated with the production of secondary metabolites, including mycotoxins. This coordination requires velvet domain proteins which are common for filamentous fungi but absent in yeast [6]. The velvet heterodimers VeA-VelB and VosA-VelB have different developmental functions. VeA-VelB heterodimer promotes sexual development whereas VelB-VosA dimer inhibits asexual differentiation. Association of the putative methyltransferase LaeA [29] with the VelB-VeA heterodimer, which makes the VelB-VeA-LaeA trimeric complex, coordinates development and secondary metabolism [6,30,31].
Comparison of the intracellular molecular mechanism of signal transduction of Fus3 MAPK of yeast and A. nidulans revealed that AnFus3 MAPK can reach the nuclear envelope in a complex with other proteins of the MAPK module, including the adaptor protein AnSte50. Only AnFus3 enters the nucleus and phosphorylates VeA, which elucidates a novel link between MAPK and velvet domain proteins that act as control elements at the interface of fungal development and secondary metabolism.

Results
The A. nidulans Fus3 MAP kinase of the mating pathway phosphorylates the velvet domain protein VeA, and VelB-VeA complex formation is reduced in Anfus3 deletion S. cerevisiae Fus3 interacts with transcription factor Ste12 that activates the mating pathway. The A. nidulans MAP kinase AnFus3 [MpkB] also controls sexual development [25,28,32]. Tagged AnFus3 recruited the transcription factor AnSte12 [SteA] by tandem affinity purification (TAP) only when the fungus was induced for sexual development but not during vegetative filamentous growth or asexual development ( Figure 1B, Table  S1). Endogenously expressed AnFus3::sGFP was functional ( Figure  S1) and immunoprecipitation of the fusion protein was able to enrich the SteA protein in a sexually induced culture (Table S1). The AnFus3-SteA interaction was further verified by bimolecular fluorescence complementation (BiFC) and was observed in fungal nuclei ( Figure 1C). This corroborates that the interaction between kinase and transcription factor is conserved from yeast to filamentous fungi.
Due to their similar roles in development and secondary metabolism [28][29][30], we examined whether AnFus3 interacts with the velvet domain proteins and LaeA. AnFus3 interacted in vivo in a BiFC assay with LaeA and subsequently with VeA, but not with VelB. In addition, AnFus3 interacted with VosA ( Figure 1D). VosA is part of the VosA-VelB heterodimer which represses asexual development [31,33]. These results suggest that distinct velvet domain proteins or LaeA may include targets of MAPK phosporylation.
AnFus3 was immunoprecipitated from vegetatively grown fungal cells as sGFP fusion protein ( Figure 2A) to identify direct substrates of AnFus3 in in vitro kinase assays. VeA expressed and purified from E. coli was the only tested protein which could be specifically phosphorylated by AnFus3, whereas bacterially produced VosA, LaeA or VelB were not phosphorylated. Further phosphorylation experiments performed with phospho-specific serine and threonine antibodies further supported that VeA was phosphorylated by AnFus3 and treatment of phosphorylated samples with lambda protein phosphatase (l-PP) resulted in loss of phosphorylation signal ( Figure 2B).
VeA bridges VelB and LaeA in the trimeric VelB-VeA-LaeA complex. We addressed whether AnFus3 activity affects complex formation. VeA protein levels ( Figure 2C) were similar in wild type and mpkB mutant strains. velB RNA was unchanged whereas laeA transcripts were downregulated as previously reported ( Figure  S2A) [28]. TAP purification of natively expressed VeA::cTAP revealed that under conditions where sexual development was normally promoted, only significantly reduced amounts of VelB and LaeA proteins were enriched by tagged VeA in the absence of MpkB ( Figure 2D, Tables S2 and S3). The MAP kinase does not affect VeA nuclear import, because the interaction of VeA with the importin KapA was not significantly affected in mpkB mutant. Consistently, nuclear import of the subunits of the trimeric VelB-VeA-LaeA complex was not affected in a mkkB mutant lacking the upstream MAP2K AnSte7 ( Figure S2B). Lack of laeA normally causes enhanced VeA and VelB expression as well as enhanced complex formation [31]. This suggests that decreased VeA-VelB association is not a result of the reduced levels of LaeA in mpkB mutants.
These results suggest that AnFus3 phosphorylates VeA in vitro and interacts with VeA in vivo. Furthermore, AnFus3 is required for enhanced association of VeA with VelB which are components of the VelB-VeA-LaeA velvet complex.
MAP2K AnSte7 is required for sexual development of A. nidulans MAPKKK (SteC) and MAPK (MpkB) are necessary for sexual development in A. nidulans [24,25]. Yeast Fus3 receives the phosphorylation signal from MAP2K Ste7. The corresponding filamentous fungus homolog has not yet been described. The ANID_03422 (mkkB) locus of A. nidulans encodes a protein, which is

Author Summary
Mitogen activated protein (MAP) kinase cascades are conserved from yeast to man to transmit an external signal to the nucleus and induce an appropriate cellular response. The yeast Fus3 MAP kinase module represents a textbook paradigm for signal transduction. The pathway is activated by external sexual hormones triggering several kinases that transmit the signal at the plasma membrane to Fus3. Phosphorylated Fus3 is released from the membrane-associated module, crosses the cytoplasm, and enters the nucleus to activate transcription factors for sexual development. We describe here the Fus3 MAPK pathway of a filamentous fungus that controls sexual development as well as secondary metabolism, which are coordinated processes in filamentous fungi. Aspergillus nidulans is able to release Fus3 as a complex from the membrane. Complexes of Fus3 can include two additional kinases and an adaptor protein, and these complexes can migrate from the membrane to the nuclear envelope where only A. nidulans Fus3 can enter the nucleus to control nuclear regulators. Revealing specific functions of cellular Aspergillus Fus3 complexes in signal transduction to control fungal development and secondary metabolism will be a fascinating future task.
conserved in different Aspergilli ( Figure S3) and has 25% identity to yeast Ste7 [34]. AnSte7 [MkkB] is also related to N. crassa MAP2K [35] and human MAP2K1 [36]. Overexpressing the corresponding mkkB gene resulted in two fold increase in the number of fruiting bodies and supported a role in sexual development ( Figure S4A-S4C). mkkB deletion mutants had a slow growth phenotype and were blocked in early sexual development, which resulted in nest-like structures containing clumps of Hülle cells (yellow arrows, Figure 3A, 3B). Hülle cells support sexual development as specialized nursing cells for the growing fruiting body [31].
AnSte7 is required for hyphal fusion as one of the initial steps of fruiting body formation. Hyphal fusion of wild type strains marked with either synthetic cytoplasmic green fluorescent protein (sGFP) or with nuclear monomeric red fluorescent protein (mRFP) resulted in hyphae with green cytoplasm and red nuclei (heterokaryon) ( Figure 3C). In contrast, a mkkB deletion strain was unable to fuse with the wild type strain. We found the same hyphal fusion defect for the steCD strain as in the mkkB mutant ( Figure 3C, 3D). This further supports that AnSte11 and AnSte7 act in a common pathway.
The analysis of putative additional functions of AnSte7 in later phases of sexual development required a by-pass of initial hyphal fusions. Therefore, heterokaryons were artificially produced by fusing protoplasts. An intact mkkB copy of the wild type strain allowed the development of mature fruiting bodies (red arrows), when wild type and mkkB mutant protoplasts were fused. In contrast, two mkkB mutants forced to form heterokaryons were impaired in fruiting body maturation and produced only early structures of development (yellow arrow, Figure 3D). This suggests several functions of MAP2K AnSte7 during sexual development presumably in concert with AnFus3.
AnSte50-Ste11-Ste7-Fus3 form a physically interacting module that is required for sexual development We determined whether the A. nidulans kinases may replace functions of its yeast counterparts. Plasmids containing Anste7 [mkkB] and Anfus3 [mpkB] genes expressed under yeast promoters were transformed into ste7 and fus3 deletion strains. mkkB and mpkB did not alleviate the defects in pheromone response of the yeast mutants ( Figure S4D). However, MpkB moderately suppressed the defects in pheromone response of a fus3 kss1 double mutant, showing that the MpkB is partially able to take over functions of the MAP kinase pair Fus3/Kss1. This suggests a partial overlap of the functions of the MAPK pathways of these two organisms.
Ste50 functions as an adaptor for membrane recruitment of Ste11 in yeast [12]. Deletion of the corresponding steD in A. nidulans caused a defect in fruiting body formation ( Figure S1A). Similar to the other MAPK mutants, steD mutant could not produce heterokaryons in outcrossings (not shown). Thus, the adaptor AnSte50 is as important for accurate fungal development as the other components of the MAPK module. A. nidulans AnSte50 was enriched by AnSte7::TAP in wild type, but not in the steCD strain indicating that AnSte11 is required for the AnSte50-Ste7 interaction ( Figure 4A, 4B). These data suggest a physical interaction of AnSte50 and two MAPK module components in a AnSte50-Ste11-Ste7 complex.
Interaction partners of AnSte50 were identified to explore the entire fungal MAPK mating module. A functional steD::ctap ( Figure  S1A, S1B) recruited the MAP3K AnSte11 and the MAPK AnFus3 but not the MAP2K AnSte7 ( Figure 4C, 4D and Table S6). This further supports that AnSte50-Ste11-Ste7-Fus3 forms a module similar to yeast Ste5-Ste50-Ste11-Ste7-Fus3 with the exception that a counterpart for the yeast Ste5 scaffold is missing in A. nidulans.
AnSte50-Ste11-Ste7-Fus3 represents an active MAP kinase module required for sexual development and secondary metabolite synthesis We analysed whether AnSte11 and AnSte7 act upstream of MAPK AnFus3. MAPK phosphorylation was monitored by a phospho-specific antibody against the MAPK Thr182XTyr184 motif. Phosphorylated AnFus3 was permanently detectable in vegetative wild type cultures ( Figure 4E). In contrast, modified AnFus3 was absent in mutants lacking AnSte11 or AnSte7, whereas the absence of AnSte12 did not change levels of phosphorylated AnFus3. In the absence of AnSte50, reduced phosphorylation of AnFus3 indicates some residual activity of the untethered AnSte11-Ste7 complex. This supports an active A. nidulans MAPK module consisting of AnSte50-Ste11-Ste7-Fus3 which controls fungal sexual development.
The role of AnSte50-Ste11-Ste7-Fus3 for secondary metabolism was examined. Impaired secondary metabolism had only been described for the mpkB mutants [28]. The mycotoxin sterigmatocystin (ST) levels were drastically reduced in the sterile steC, steD, mkkB, or mpkB mutants whereas ST levels in the sterile steAD [AnSte12] were similar to wild type ( Figure 4F-4G). Similarly, the expression of the biosynthesis genes for ST (stcU) and terrequinone (tdiA and tdiB), and the expression of laeA and the transcription factor encoding aflR, both required for expression of secondary metabolite genes, were distinctly reduced in each mutant of the MAPK module ( Figure 4H). These data corroborate that active , LaeA and velvet domain proteins. Germination of spores leads to tubelike vegetative filaments (hyphae) which become competent for environmental signals after at least 12 hours of growth. Exposure of developmentally competent hyphae to light (or aeration) leads to asexual development (conidiophores and asexual spores [conidia]) in 24 hours. VosA-VelB inhibits asexual differentiation. Incubation in dark (96 hours) induces the sexual cycle with sexual fruiting bodies (cleistothecia) which are nursed by globose Hü lle cells. LaeA is required for Hü lle cell formation. VelB-VeA supports sexual development together with AnSte12 [SteA]. The VelB-VeA-LaeA trimeric complex coordinates development with secondary metabolism. Co; conidia, S; stalk, Cl; cleistothecium, Hc; Hü lle cells. (B) A silver stain treated 5-14% gradient SDS polyacrylamide gel of AnFus3 [MpkB]::cTAP from vegetatively, asexually (on plates, under light) and sexually (on plates, in the dark) grown cultures at 30uC for 20 hours. Identified proteins from the excised lanes (Table S1)  In vitro phosphorylation of VeA by AnFus3 in a radioactive kinase assay. Left panels; autoradiograph of dried SDS gel run for phosphorylation reactions (30 ml of total 45 ml reaction volume), Coomassie stain of the proteins from phosphorylation reaction (10 ml of total 45 ml reaction). VeA protein in AnFus3 reaction tube is shown with red rectangle. Right panels; ponceau staining of the immunoprecipitated (immobilized to the GFP trap sepharose) AnFus3::sGFP and only sGFP protein. Immunodetection of the fusion protein and free sGFP by the a-gfp. (B) Confirmation of specific VeA phosphorylation by a non-radioactive method. All recombinant proteins (10 mg each) were treated with both AnFus3 and GFP. AnFus3 treated samples were additionaly incubated with the lambda protein phosphatase (l-PP). Proteins were immunodetected by a-Penta His, a-GST. After AnFus3 treatment, VeA showed a 3-5 kDA molecular weight shift (red arrow) that disappeared after L-PP treatment. Only VeA treated with MAPK was recognized by P-ser/thr specific antibody. (C) Protein levels of VeA in the wild type and mpkB mutant background. VeA::cTAP signals were normalized to the internal actin levels. VeA protein levels did not change in the absence of MpkB. (D) Reduced velvet complex formation in the mpkB mutant. The VeA-associated proteins from the cultures of the wild type and mpkBD strains grown in the darkness sexually at 30uC for 20 hours. Three independent experiments were performed and the associated proteins were identified. The ratio of the peptides from VelB and LaeA to the VeA protein drastically reduced in the MAPK mutant, whereas alpha importin KapA interaction slightly increased. Black bars represent the standard error. *LaeA was only found in one of the three purifications in mpkBD strain, thus no error bar is assigned. doi:10.1371/journal.pgen.1002816.g002 AnSte50-Ste11-Ste7-Fus3 MAPK is not only required for sexual development but also for secondary metabolite production.
The components of the fungal MAPK module exhibit distinct localization patterns at hyphal tip, nuclear envelope, and septa The yeast mating MAPK module transmits a signal from the plasma membrane to the nucleus by releasing MAPK Fus3 from the Ste5 scaffold at the membrane [13,37]. We analysed how the signal is transmitted through the filament of A. nidulans to nuclear factors as AnSte12 or VeA. Time course immunoblotting ( Figure S1C) showed that AnFus3 was constantly expressed during development. The mkkB mRNA for the upstream MAPKK was also present throughout all stages ( Figure S5C). The corresponding protein AnSte7::sGFP was present in vegetative as well as in the initial phases of asexual and sexual development, but decreased afterwards ( Figure S5C). Similarly, the AnSte50::sGFP ( Figure S1C) seems to be degraded because the protein disappeared during mid and late asexual development.
Confocal spinning disc microscopy revealed that functional AnSte7::sGFP fusion protein expressed under native locus promoter was localised during early phase of growth throughout the cytoplasm, but never found in the nucleus (not shown). After becoming competent for differentiation (16 hours after germination), AnSte7::sGFP accumulated not only at the hyphal tip but also at the plasma membrane and at the septa of hyphae or spore forming cells (white arrows in Figure 5A). The AnSte7 signal was also present on the nuclear envelope. The AnSte7 localization pattern did not change in the absence of the MAP3K AnSte11 (not shown). Like AnSte7, a functional Ste50::sGFP fusion never entered the nucleus. AnSte50 was cytoplasmic and accumulated at later stages of vegetative growth at the hyphal tip, the septa of spore forming cells, the plasma membrane and the nuclear envelope (arrows in Figure 5B, 5D).
A functional AnFus3::sGFP expressed under the native promoter accumulated at the hyphal tip and was as well present in the cytoplasm as in the nucleus in vegetative and spore forming cells ( Figure 5C, 5D). This suggests a dynamic and complex distribution of MAPK module subunits from the fungal membrane to the nucleus. It also revealed that the MAPK AnFus3 is as yeast Fus3 the only subunit with the potential to enter the nucleus.
The entire MAPK module colocalizes and interacts at hyphal tip and nuclear envelope  Figure 6). Most of the GFP signals of AnSte7 and Ste50 merge with the RFP signal of MpkB at the fungal tip, the plasma membrane and at the nuclear envelope where they might form dynamic protein complexes. Exclusively at hyphal tips we found two types of co-localizations of kinase pairs. In addition to direct co-localizations, similar to plasma membrane or nuclear envelope, there were extended colocalization patterns at the hyphal tip. This could reflect that a fraction of kinases is localized in vesicles at the hyphal tip.
Bimolecular fluorescence complementation (BiFC) [38,39] was applied to examine whether there are direct transient in vivo interactions between AnSte7 and Fus3, which could not be found by TAP purification ( Figure 7B). Similar to the yeast localization of the Ste5-Ste11-Ste7-Fus3 MAPK module at the membrane,  AnSte11-Ste7 and AnSte7-Fus3 interacted at the plasma membrane and also at the hyphal tip ( Figure 7). There was an additional strong interaction of AnSte11-Ste7 at septa which border cellular segments as well as at septa of spore forming cells and spores ( Figure 7C-7E). Quantification of the fluorescence intensity from the bright enhanced yellow fluorescent protein (EYFP) spots of AnSte11-Ste7, Ste7-Fus3 pairs revealed that they emit upto 10 fold more yellow fluorescence than the single EYFP molecules ( Figure S6), suggesting that the kinase pairs form multimeric complexes.
Consistently to the yeast situation, the transcription factor AnSte12 as well as fungus specific factors VeA and LaeA specifically interacted with the MAPK Fus3 in the nucleus ( Figure 1C, 1D). AnSte50 also interacted with the kinases at the plasma membrane and hyphal tip ( Figure 8). Only AnSte11-Ste7 strongly interacted at the septa but there was hardly any interaction between AnSte7-Fus3 or between the AnSte50 and any of the kinases at the septa ( Figure 7C and 7D, Figure 8D-8F).
The entire MAPK module components migrate to the nuclear envelope to deliver AnFus3 into the nucleus Yeast Ste7-Ste5-Fus3 migrates to tips of mating projections in pheromone treated cells. Only Fus3 travels to the nucleus upon activation by Ste7 [13]. A. nidulans is a homothallic fungus, which does not require a mating partner. Time lapse images revealed that MAPK module components AnSte7 and Ste50::sGFP can move within the fungal cell along the membrane. During the cellular movements, these molecules shortly touched the membrane then hit the nucleus. Sometimes, fusion protein moved back after contacting the nucleus in the opposite direction. (Figure 9A and 9B, Videos S1 and S2).
The dynamics of the protein interactions of the BiFC expressing strains were further analysed by time lapse movies (Videos S3, S4,  S5, S6). The AnSte7-Fus3 pair moved together along the plasma membrane (Video S3, Figure S7A) towards the first nucleus, then they advanced to the next nucleus while some other spots did not move distinctly. Likewise, AnSte50-Fus3 complexes left one nuclear envelope, touched the membrane and moved to the next nucleus (Video S4, Figure S7B). Similar movements were also observed for other complexes of the MAPK module (not shown). AnSte11-Ste7 can dissociate from the plasma membrane, cross the cytoplasm and reach the nuclear envelope (Videos S5 and S6, Figure S7C, S7D).
The major difference to the yeast situation is that the MAPK module of A. nidulans travels from the outer border of the fungal cell through the cytoplasm to the nuclear envelope. The AnSte7-Fus3 pair as well as pairs of AnSte50 with all three kinases interacted at the nuclear envelope ( Figure 7 and Figure 8). These data suggest significant differences in the molecular mechanism how a MAPK signal is transmitted in yeast in comparison to a filamentous fungus.

AnSte50 is required for efficient membrane attachment of MAPK complexes
The interactions of the AnSte11-Ste7 and AnSte7-Fus3 complexes were examined in steDD strain to examine AnSte50 function for cellular location of the module. The interaction of the three kinases at the plasma membrane of wild type (Figure 7) was abolished for AnSte11-Ste7 and drastically reduced for AnSte7-Fus3 in the steD mutant ( Figure 9C, 9D). Plasma membrane localizations of the AnSte7 and Fus3::sGFP fusions were also reduced in the steD mutant (not shown). Contrastingly, the localization of the entire module at the hyphal tip or for the partial module AnSte11-Ste7 at the septum seems to be mediated by a mechanism which is largely independent of AnSte50.
These data suggest that AnSte50 supports association of the A. nidulans MAPK module with the plasma membrane but it does not affect the hyphal tip and septum localizations.

Discussion
We describe here the A. nidulans Fus3 MAPK module which is involved in sexual development and the control of secondary metabolism and releases AnFus3 into the nucleus. Our data suggest a provocative additional hypothesis: AnFus3 is able to travel along the membrane and to cross the cytoplasm to the nuclear envelope in complexes with AnSte7 MAP2K, AnSte11 MAP3K and the adaptor protein AnSte50. In the nucleus AnFus3 interacts with transcription factor AnSte12 for sexual development. The additional interaction of AnFus3 with VeA or yet unidentifed targets may promote VeA-VelB formation which is required for coordinated development and secondary metabolism ( Figure 9E). The A. nidulans Fus3 MAP kinase module is preferentially assembled at distinct intracellular locations, such as the hyphal tip, the septa, the plasma and nuclear membranes. Membrane localisation of the module is presumably relevant to perceive external signals as in yeast. Sexual development is defective when membrane localization of the module is impaired as in strains without intact AnSte50. Tip localisation could be important for hyphal fusions and cell-cell contacts. MAPKK AnSte7 and MAP3K AnSte11 but not other components interact at septa suggesting additional phosphorylation functions at septa independent of AnFus3. Corresponding mutants displayed strong deformations in the septa between developing asexual spores and spore forming cells but did not show any abnormal septation pattern in vegetative hyphae (not shown). This suggests a possible additional link between kinases of the module and regulators of asexual development.
Intracellular distances in a filamentous fungus are significantly larger than in yeast. Several steps can be distinguished for signal transduction from surface to nucleus of A. nidulans. (i) From hyphal tip to plasma membrane: AnSte50 is primarily required for efficiently anchoring the MAPK module to membranes, but not to hyphal tips. AnSte50 might also contribute like in yeast to Ste11 MAP3K activation. The essential function of AnSte50 for signal transduction is supported by the defect of sexual development and lack of AnFus3 phosphorylation in a steD mutant. The AnSte50 independent localization at the hyphal tip suggests an additional yet unknown anchoring function for the AnFus3 module at the hyphal tip. The anchoring mechanism could include small membrane bound vesicles at the Spitzenkörper which could explain some of our localization results ( Figure 6B, Figure 7, Figure 8).
The lack of AnSte11 did not cause any changes in the subcellular localization of AnSte7, indicating that AnSte11 is not required for proper AnSte7 localization. The lack of AnSte50 had a drastic effect on the localization of MAPK module complexes. AnSte50 interacts with all components of the MAPK module and might provide a binding platform for the other MAPK components which even works when AnSte11 is absent ( Figure 8).
(ii) In yeast Fus3 dissociates from the Ste5 tethered pheromone pathway module and enters into the nucleus [13]. Transport of the AnFus3 in the AnSte50-Ste11-Ste7 complex (or subcomplexes) to the nuclear envelope as additional signal transmission step in A. nidulans might secure that AnFus3 can be kept active over larger distances until it finally reaches the nucleus. It will be interesting to analyse phosphorylation states of kinases at different cellular locations during signal transduction.
(iii) Import of AnFus3 from nuclear envelope into nucleus: AnFus3 presumably dissociates from the kinase module at the nuclear envelope in a mechanism wihich is unknown. After entry into the nucleus, AnFus3 interacts with AnSte12, and presumably phosphorylates it. AnFus3 phosphorylates the velvet protein VeA, which efficiently associates with VelB and LaeA. It is yet unclear whether there are additional AnFus3 targets which support VelB-VeA complex formation. VelB-VeA then contributes with AnSte12 to sexual fruiting body development and the trimeric VelB-VeA-LaeA concomitantly promotes expression of distinct genes for secondary metabolites ( Figure 9E). These include the production of the mycotoxin sterigmatocystin or antitumor agent terrequinone but not the antibiotic penicillin synthesis.
The MAPK module of A. nidulans is presumably involved in integrating multiple signals and enabling an adequate cellular response. Oxylipins represent currently the only known pheromones of Aspergilli but the receptors are unknown [40]. In yeast nitrogen starvation induces the same kinase module as pheromones, and part of the components are also involved in response to osmotic stress. It is likely that the AnSte50-Ste11-Ste7-Fus3 and the septal AnSte11-Ste7 modules have additional targets other than AnSte12 and VeA, which remain to be identified.
An interesting open question is whether other organisms also transport their Fus3 MAPK counterpart together with the entire module from surface to nuclear envelope. This results in questions about transport control points and module attachment sites on the nuclear envelope where future work in A. nidulans could deepen insights into the molecular mechanism of information transfer through the cell.

Manipulation of nucleic acids
Circular and linear DNA molecules were created based on the standard recombinant DNA technology protocols in detail [30]. Plasmids and oligonucleotides applied and constructed in this study are given in Table S8 and Table S9.

Hybridization techniques and analysis of nucleic acids
Southern and Northern hybridizations were carried out as explained in detail [30,43] according to protocols [48,49].

Immunoblotting
Immunoblotting experiments for recognition of GFP, TAP fusion, VeA, and actin in protein extracts was performed according to described protocols [31]. a-phospho 44/42 (4377, CELL SIGNALING TECHNOLOGY INC) was used for detection of the phosphorylated AnFus3 [MpkB]. For the detection of the phosphorylated proteins, a-phosphoserine/threonine (ab17464, ABCAM) was employed. Manufacturers protocols were followed for incubation times and buffer applications of phosphospecific antibodies.

Expression of recombinant proteins
Proteins were expressed in Rosetta 2 (DE3) using ZYM5052 [50] media supplemented with 30 mg/ml Chloramphenicol and 100 mg/ml Ampicillin (GST-LaeA91) or 30 mg/ml Kanamycin (Velvet proteins) at 16uC. Cells were harvested by centrifugation, resuspended in lysis buffer (30 mM HEPES pH 7.4, 400 mM NaCl, 30 mM Imidazol) and lysed by passing through a Microfluidics Fluidizer at 0.55 MPa. The lysate was cleared by centrifugation at 300006g for 30 minutes. His-tagged proteins were purified with a 5 ml NiNTA-Sepharose (GE HEALTHCARE) and GST-tagged LaeA91 with a 5 ml GSH-Sepharose (GE HEALTHCARE) column connected to an Ä KTA Prime chromatography system. After washing with 10 column volumes with lysis buffer, proteins were eluted with elution buffer plus 400 mM Imidazol or 30 mM reduced Glutathione. Velvet proteins were desalted with a HiPrep Desalting 26/10 column (GE HEALTHCARE) into storage buffer (10 mM HEPES pH 7.4, 400 mM NaCl). GST-LaeA91 was cleaved with PreScission Protease at 4uC for 16 h and further purified by gel-filtration using a Superdex 200 26/60 and a final 5 ml GSH-Sepharose column both equilibrated in gelfiltration buffer (10 mM HEPES pH 7.4, 150 mM NaCl). All proteins were shock-frozen in liquid nitrogen and stored at 280uC until further use.

Protein immunoprecipitation
In order to immunoprecipitate GFP fusion proteins, protein crude extracts were prepared from vegetatively grown cultures. 100 ml GFP-Trap sepharose (CHROMOTEK) was washed twice with 1 ml protein extraction buffer (50 mM Tris pH 7.5, 100 mM KCl, 10 mM MgCl 2 , 0.1% NP40, 10% Glycerol, 20 mM b-glycerophosphate, 2 mM Na 3 VO 4 , 5 mM NaF, 0.5 mM PMSF, 1 mM benzamidine, 1 mM EGTA, 1 mM DTT). 20 ml (150 mg total) protein crude extract was incubated with 100 ml GFP-Trap sepharose (CHROMOTEK) at 4uC for 2 hours on a rotating platform. Afterwards, sepharose-extract mixture was centrifuged at 4000 rpm at 4uC for 1 min. Crude extract was removed with a 5 ml pipette. The sepharose was washed twice with 20 ml of protein buffer and centrifuged at 4000 rpm at 4uC for 1 min. This step was repeated one more time. Finally, 1 ml of protein buffer was added and the sepharose was resuspended. Each of the 200 ml sepharose buffer mixture was transferred into 1.5 ml eppendorf cups and centrifuged at 4000 rpm at 4uC for 1 min and supernatant was removed. Immunoprecipitated proteins were washed three times with 1 ml kinase reaction buffer (KRB; 20 mM Tris pH 7.5, 10 mM MgCl 2 , 1 mM DTT, 1 mM benzamidine, 1 mM Na 3 VO 4 , 5 mM NaF, 0.1 mCi [ 32 P]-ATP).

In vitro phosphorylation and dephosphorylation assay
In vitro phosporylation assay was performed with modifications according to protocol given in [51]. For in vitro phosphorylation experiment, 30 ml KRB, containing 0.1 mCi [ 32 P]-ATP and 10 mg recombinant protein were added to the sepharose beads and incubated at 30uC for 35 minutes with the periodic resuspensions in every five minutes. Afterwards, reaction tubes were centrifuged at 4000 rpm at R/T for 1 min and supernatants containing phosphorylated proteins were transferred into new eppendorf cups. Supernatans and sepharose containing immunoprecipitated proteins were mixed with 36 protein loading dye (30 ml supernatant and 15 ml loading dye) and incubated at 95uC for 10 min. 30 ml of the supernatant fraction was run on 4-15% gradient SDS gel that was dried for 2 h and exposed to Kodak X-omat film for 5 hours. 10 ml of the reaction was used for visualization of the proteins with coomassie staining. 2 ml of sepharose was used for immunoblotting and ponceau staing for validation of equal immunoprecipitated target protein (MpkB or GFP). For non-radioactive kinase experiments, same KRB buffer containing 5 mM ATP was used. Supernatants were treated with 1000 units lambda protein phosphatase (NEW ENGLAND BIOLABS) in the presence of 1 mM MnCl 2 at 30uC for 1 hour. Samples were added with 36 loading dye and boiled at 95uC for 10 min. 3 ml of the samples were used for immunoblotting.

Tandem Affinity Purification (TAP) protocol and LC-MS/ MS protein identification
For the TAP purification of the MkkB, MpkB, SteD, and VeA interacting proteins and further LC-MS/MS identification previously published protocols were applied [30].

Confocal spinning disc and fluorescence microscopy
A. nidulans strains expressing various fluorescence proteins (EYFP/sGFP/mRFP) were inoculated in the 8-well borosilicate coverglass system (NUNC) containing the liquid minimal medium. Widefield fluorescence photographs were taken with an AXIOVERT OBSERVER. Z1 (ZEISS) microscope equipped with a COOLSNAP ES2 (PHOTOMETRICS) digital camera. CSU-X1 A1 confocal scanner unit (YOKOGAWA) connected with QUANTEM:512SC (PHOTOMETRICS) digital camera was used for laser microscopy. The SLIDEBOOK 5.0 software package (INTELLIGENT IMAGING INNOVATIONS) was used for fluorescence and laser confocal image and movie recording as well as productions. We defined signals as plasma-membrane localized if we found the signals that are at the border of the silhouette of the fungal cell or even surmount the fungal cell; similarly, we defined signals as nucleus-associated when we found multiple signals at the border of the nuclear silhouette.

Quantification of the YFP fluorescence
The EYFP protein was purified by using GFP-Trap as described for GFP protein. EYFP molecules were allowed to attach to poly-L-lysine coated coverslips for 10 minutes, in PBS buffer. Fungal cultures were grown as described above. The preparations were imaged using a SP5 TCS STED microscope (LEICA MICROSYS-TEMS), under 514 nm excitation (provided by an Argon laser), using a 1006oil-immersion objective (1.4 NA, LEICA). The images were processed by a custom-written routine in Matlab (THE MATHWORKS INC.). Briefly, the spots were identified by the application of an automatic threshold based on the intensity of the background. We then used Gaussian fits to the spots to determine their intensity, and to correct for the background intensity, which provided the baseline of the fits.

Analysis of secondary metabolites
Extraction of sterigmatocystin (ST) and thin layer chromatography (TLC) was carried out as given in detail [43]. Penicillin levels were determined as published previously [52]. [mpkB] expressed under yeast STE7 or FUS3 promoters in a centromeric self-replicating plasmid. These constructs were expressed in the respective fus3, ste7 and fus3/kss1 double mutants. Strains were grown in the presence of 15 mg alpha factor given on the paper discs at 30uC for 3 days. Alpha factor in wild type (empty plasmid) and complementation strains (STE7 in ste7 mutant, FUS3 in fus3 mutant, FUS3 in fus3/kss1 mutant) results in a strong growth inhibition (halo). ste7 and fus3/kss1 mutants do not show any response to the pheromone treatment. fus3 mutant exhibits a reduced response (cloudy halo). AnSte7 and Fus3 do not remediate the halo phenotype of the ste7 and fus3 mutants. mpkB cDNA partially restores the pheromone response of the fus3/kss1 double mutant. (TIF)  Figure 7 for examples). The intensities are comparable, although the values tend to be higher for AnSte11-Ste7 (blue) than for AnSte7-Fus3 (red). To obtain an estimate of the number of molecules in the complexes, these values were compared to the intensity of single EYFP molecules attached to coverglasses (green). (B) The bar graphs indicate the average values for the intensities (obtained from the same datasets as in A). The intensity of the complexes is ,9-fold (AnSte7-Fus3) or ,10-fold (AnSte11-Ste7) higher than that of single EYFP molecules, suggesting the presence of 9 to 10 molecules in a complex. The bars show the mean and standard error; 50-300 spots were analyzed for each condition.    Video S1 Time-lapse analysis of the subcellular movements of the AnSte7-GFP fusion along the fungal cells. Individual focal planes were captured with a spinning disc confocal microscope at 2 min intervals (total 26 min). AnSte7 protein moves in an internuclear manner. The nuclei were visualized by mRFP::His-tone2A fusion protein. Green spot leaves the first nucleus and shortly touches the plasma membrane (zigzag movement) and sticks to the envelope of the next nucleus. Some spots of the AnSte7 are static (immobile dot at the hyphal tip). The video is presented at a rate of 5 frames/second.

(MOV)
Video S2 Time-lapse capture of the subcellular movements of the AnSte50-GFP fusion within the fungal cell. Single focal planes were captured at 2 min intervals (total 22 min). AnSte50 protein moves to the nucleus (red), hits the nuclear envelope or nucleus and moves in a retrograde direction. Due to the single focal plane, spot disappears at 6 min. After 6 min, spot movement can be tracked again. The video speed is 5 frames/ second.

(MOV)
Video S3 Time-lapse analysis of the subcellular movements of the AnSte7-AnFus3 complexes (yellow dot) along the fungal hypha. AnSte7-Fus3 complexes move in a retro and anterograde direction in comparison to the hyphal tip (upper left). These complexes move between the nuclei, which were visualized by mRFP::Histone2A fusion. Single focal layer images were captured at 1 min intervals (total 17 min). The video was produced at a setting 5 frames/second.

(MOV)
Video S4 Time-lapse analysis of the subcellular dynamics of the AnSte50-AnFus3 complexes (yellow dot) in the fungal hypha. Single focal pictures were taken at 2 min intervals (total 58 min, 48 min is shown). This movie shows the movement of the AnSte50-AnFus3 complexes between two nuclei. Protein complexes (yellow dot) leave the first nucleus (visualized by mRFP::Histone2A fusion) and move to the second one. While moving to the second one, the complexes slightly touch the plasma membrane. Video is presented by using the setting 5 frames/ second.

(MOV)
Video S5 Retrograde translocation of the AnSte11-Ste7 protein complexes along the fungal hypha. AnSte11-Ste7 complexes move backwards from the hyphal tip. They leave the membrane and touch the nucleus. They also accumulate at septa (faint immobile yellow dot). Single focal planes were captured at 2 min interval (total 54 min, 28 min is shown). The video is presented at the speed of 5 frames/second. (MOV) Video S6 Time-lapse analysis of the migration of AnSte11-Ste7 complexes from the plasma membrane to the nucleus. Membranetethered complexes (yellow spots) slowly move to the nucleus (red). The movie was captured at 1 min interval (total 39 min). 5 frames/second. (MOV)