Assembly of a heptameric STRIPAK complex is required for coordination of light-dependent multicellular fungal development with secondary metabolism in Aspergillus nidulans

Eukaryotic striatin forms striatin-interacting phosphatase and kinase (STRIPAK) complexes that control many cellular processes including development, cellular transport, signal transduction, stem cell differentiation and cardiac functions. However, detailed knowledge of complex assembly and its roles in stress responses are currently poorly understood. Here, we discovered six striatin (StrA) interacting proteins (Sips), which form a heptameric complex in the filamentous fungus Aspergillus nidulans. The complex consists of the striatin scaffold StrA, the Mob3-type kinase coactivator SipA, the SIKE-like protein SipB, the STRIP1/2 homolog SipC, the SLMAP-related protein SipD and the catalytic and regulatory phosphatase 2A subunits SipE (PpgA), and SipF, respectively. Single and double deletions of the complex components result in loss of multicellular light-dependent fungal development, secondary metabolite production (e.g. mycotoxin Sterigmatocystin) and reduced stress responses. sipA (Mob3) deletion is epistatic to strA deletion by supressing all the defects caused by the lack of striatin. The STRIPAK complex, which is established during vegetative growth and maintained during the early hours of light and dark development, is mainly formed on the nuclear envelope in the presence of the scaffold StrA. The loss of the scaffold revealed three STRIPAK subcomplexes: (I) SipA only interacts with StrA, (II) SipB-SipD is found as a heterodimer, (III) SipC, SipE and SipF exist as a heterotrimeric complex. The STRIPAK complex is required for proper expression of the heterotrimeric VeA-VelB-LaeA complex which coordinates fungal development and secondary metabolism. Furthermore, the STRIPAK complex modulates two important MAPK pathways by promoting phosphorylation of MpkB and restricting nuclear shuttling of MpkC in the absence of stress conditions. SipB in A. nidulans is similar to human suppressor of IKK-ε(SIKE) protein which supresses antiviral responses in mammals, while velvet family proteins show strong similarity to mammalian proinflammatory NF-KB proteins. The presence of these proteins in A. nidulans further strengthens the hypothesis that mammals and fungi use similar proteins for their immune response and secondary metabolite production, respectively.


Introduction
Signaling pathways that regulate morphological and physiological processes in response to stimuli are often highly conserved throughout eukaryotes, signifying their importance. Striatin is one of the regulatory proteins proposed to act as a signalling hub for the control of many cellular processes including development, cellular transport and signal transduction [1,2]. It forms a scaffolding platform to build the striatin-interacting phosphatase and kinase (STRIPAK) complex which is a large multimeric protein complex highly conserved in eukaryotes [3]. The STRIPAK complex influences mammalian cell size, morphology and migration [1]. It also plays a role in the polarisation of the golgi apparatus and is implicated in the process of mitosis through tethering vesicles of the golgi to the nuclear membrane and centrosomes [4].
The STRIPAK complex in the fruit fly Drosophila melanogaster acts as a negative regulator of the Hippo signaling pathway [10]. In Saccharomyces cerevisiae, the homologous complex is termed the Far complex which is implicated in cell cycle arrest and acts as an antagonist towards target of rapamycin complex 2 (TORC2) signaling [11,12]. In the fission yeast Schizosaccharomyces pombe, the STRIPAK complex is known as the SIP complex (Septation

The AnSTRIPAK complex is a heptameric protein complex
To reveal and understand the molecular interaction network of StrA (AN8071), a fully functional StrA-TAP fusion expressed from its native locus under the control of the endogenous promoter was used for tandem affinity purification (TAP) and liquid chromatography-mass spectrometry (LC-MS/MS) (Fig 1A, S1-S6 Tables). Six proteins associated with StrA, termed striatin interacting proteins (Sips), were identified from vegetative cultures grown for 24 hours: AN6190/SipA, AN1010/SipB, AN6611/SipC, AN4632/SipD, AN0164/SipE (PpgA) and AN4085/SipF. For consistency, SipE and PpgA refer to the same protein within the text. StrA was expressed at higher levels during early stages of asexual and sexual development in comparison to vegetative growth (S1 Fig). Therefore, TAP-MS/MS was performed from light and dark induced cultures (6, 24, 48h). StrA interacted with the same set of proteins SipA-F during early sexual and asexual development (6h). However, it recruited only SipA at late developmental time points (24, 48h) (Fig 1A).
With the exception of SipB, all other interacting partners of StrA are conserved components of the STRIPAK complex in fungi (S7 Table). SipA is an ortholog of human and fungal Mob3 (Phocein) with 480 amino acids (Fig 1B and 1C). Mob3p is a kinase co-activator and a part of the STRIPAK complex in humans, fruit flies and filamentous fungi [15]. Interestingly, The Cterminus of SipB (444 aa) contains a suppressor of IKKε (SIKE)-like domain (S2 Fig), an ortholog of which has only recently been shown to exist in S. macrospora [17]. SIKE-like proteins contain a coiled-coil domain which is conserved in fungi. Alignment of two human SIKE isoforms with the c-terminus of A. nidulans SipB indicates 46% similarity between the proteins (S2 Fig).
Since SipA interacted with StrA during all developmental stages, expressions of the functional Sip-GFP fusions (S3 Fig) were monitored during vegetative, asexual and sexual stages ( Fig 1D). The catalytic subunit SipE (PpgA), but not the regulatory subunit of phosphatase (SipF) was used for expression studies since sipF deletion was lethal. A 79kDa SipA-GFP and a 76kDa SipB-GFP fusion were present during almost all developmental time points. A 145kDa SipC-GFP fusion was poorly expressed during developmental stages. A 107kDa SipD-GFP fusion was constantly expressed at all stages except for late sexual development (24 and 48h). Interestingly, the 63kDa SipE-GFP fusion was only present during vegetative growth and degraded and disappeared at both asexual and sexual stages.
In order to define the core STRIPAK complex more precisely, reciprocal TAP-MS/MS was performed. TAP of SipA recruited StrA along with SipB to SipF (Fig 1E). Similarly, TAP of SipB, SipC, SipD or SipE also recruited all members of the complex (S8-S12 Tables). These interactome data clearly underline that a heptameric STRIPAK complex made of a striatin StrA, a Mob3 kinase ortholog SipA, a SIKE-like protein SipB, a STRIP1/2 ortholog SipC, an SLMAP ortholog SipD, and the phosphatase subunits SipE (PpgA) and SipF exists in A. nidulans. . A strain expressing a functional StrA-TAP fusion was grown in liquid GMM (grown vegetatively, i.e. neither sporulation or the sexual cycle are initiated in liquid cultures) (24h). For induction, cultures were grown vegetatively for 20h and shifted on plates to be induced for both light (6 and 24h) and dark (6, 24 and 48h) development by growing them further at 37˚C. Therefore, light and dark cultures also represent SM production phase. TAP-MS was performed as described in materials and methods. Unique peptides numbers are given (S1-S6 Tables). (B) Domain structure of the STRIPAK complex members: StrA, SipA, SipB, SipC, SipD, SipE (PpgA) and SipF. N-C; N-and C-

The AnSTRIPAK complex is required for proper development and light response
To understand the roles of the AnSTRIPAK complex in fungal development, individual sip deletions, sip;sip double deletions and combinations of sip deletions with strAΔ were created (S4 Fig) and subjected to developmental tests (Fig 2). Loss of sipB, sipC, sipD and sipE all resulted in similar phenotypes to that seen in the strAΔ strain, characterized by slower growth rate, reduced conidiation and lack of fruit bodies. sipE (ppgA), encoding one of the two phosphatase 2A catalytic subunits of A. nidulans was shown to influence growth together with more than twenty other phosphatases [27]. All attempts to delete sipF encoding PP2A-A regulatory subunit failed, suggesting that the gene is essential for viability of A. nidulans as it is in N. crassa [15].
The defects were complemented by introducing the corresponding genes into the deletion strains (Fig 3). Furthermore, sipΔ/sipΔ double deletions and sipΔ/strAΔ double deletions were similar to single deletion strains. Surprisingly, sipAΔ showed an opposite phenotype with significantly increased radial growth and two-fold more conidiation than the WT and production of normally shaped fruit bodies, which were devoid of ascospores (Fig 2). Moreover, strAΔ/ sipAΔ phenocopied sipAΔ, suggesting an epistatic effect of sipA over the strA. In contrast, sipA was not epistatic to sipB to sipE since sipA double deletion combinations with other sip genes behaved similar to single sip deletions, showing that sipA is only epistatic to StrA yet not the other members of the AnSTRIPAK complex. These results show that almost all members of the AnSTRIPAK complex are equally important for growth and light-dependent development except for SipA. However, SipF is likely essential for the viability of A. nidulans.

The AnSTRIPAK complex is involved in responses to various environmental stress factors
The STRIPAK complex mutants showed developmental defects, suggesting an essential role of the complex in signal transduction of developmental processes. In order to see whether the AnSTRIPAK complex is also involved in stress responses, mutants were subjected to various stress conditions (Fig 3). The radial growth of all single and double deletions except for sipAΔ were significantly reduced in oxidative stress (H 2 O 2 , Menadione) and cell wall stress (Congo Red) media in comparison to WT (Fig 3A-3C). sipAΔ was more resistant to oxidative and cell wall stressors than the WT. However, under stress conditions sipAΔ did not show epistatic effects over the strAΔ since the sipA/strA double mutant was as sensitive as the strA single deletion and the double deletions of strA and sip genes to both stress conditions. Strains were also monitored to see how they cope with DNA damage, amino acid starvation (3-AT), caffeine and osmotic stress. Similar to oxidative and cell wall stress, sipAΔ displayed more robust vegetative growth than the WT under all tested conditions. strA double deletions with sip genes, interestingly including the strA/sipA double mutant, were extremely sensitive terminus, (aa) Number of amino acids. WD (Trp-Asp) repeats, Mob3: Monopolar spindle one-binder protein (Phocein); N1221: acidic domain with possible transmembrane domains, DUF3402; FHA: Forkhead-associated domain, HEAT repeats. (C) Orthologs of STRIPAK complex from yeast to human. The table was mainly adapted from yeast [6] and A. nidulans STRIPAK orthologs were highlighted. Detailed accession numbers are given in S7 Table. (D) Expression patterns of SipA to SipE (PpgA) proteins during fungal development. Vegetatively grown cells (24h) were shifted onto plates and induced for asexual (6, 12, 24h) and sexual (6, 12, 24, 48h) development. Sip-GFP fusions (100 μg total protein) were used for time course immunoblotting. An α-GFP antibody was used to detect fusions and α-SkpA shows equal loading in each lane. (E) Confirmation of components of the STRIPAK complex by reciprocal TAP-LC-MS/MS. Functional Sip-TAP fusions, grown vegetatively for 24h, were subjected to TAP-LC-MS/MS. TAP tagged proteins are given at the top of the table and proteins copurified are given on the left-hand side of the table. Unique peptides are given as numbers (S8-S12 Tables). https://doi.org/10.1371/journal.pgen.1008053.g001 Fungal STRIPAK complex assembly All of the deletion strains excluding the sipA deletion were slightly sensitive to osmotic stress induced by NaCl.
These data clearly indicate that the lack of AnSTRIPAK complex results in drastic problems in combating various types of stressors. Furthermore, the epistatic effect of SipA over the StrA deletion is abolished in the presence of stress conditions, suggesting an interplay between SipA and StrA in regulating stress responses.

The AnSTRIPAK complex controls expression of developmental regulators and reactive oxygen species (ROS) scavenging enzyme genes
All mutants of the AnSTRIPAK complex, except for sipAΔ, exhibited drastic changes in asexual and sexual development, both of which are regulated by a cascade of transcription factors. We consequently determined the effects of our deletions on expression of these transcription factors by qRT-PCR. Expression of major transcription factors that control conidiophore (abaA, brlA) and sexual (nsdD, steA) development were significantly reduced in the STRIPAK mutants except for sipAΔ (Fig 4A and 4B) which caused 3-fold higher brlA, abaA and 2-fold nsdD, steA expression. This increase was consistent with the increased asexual development of the sipA mutant. AnSTRIPAK complex mutants were sensitive to oxidative stress. Therefore, expression of two ROS scavenging enzyme encoding genes, catalase catC and superoxide dismutase sodB were monitored in the mutants. Expression of both genes was significantly decreased in all AnSTRIPAK mutants except for sipAΔ (Fig 4C), where they were slightly upregulated, which was consistent with the higher resistance of this strain to stressors when compared with WT. These expression data demonstrate that the AnSTRIPAK complex is required for the balanced expression of developmental regulators and ROS scavenging enzyme genes.
The AnSTRIPAK complex plays a key role in regulation of secondary metabolite production A defect in fungal development particularly in fruit body formation (sexual development) is often associated with changes in SM production. All members of AnSTRIPAK complex are involved in growth and fruit body formation, excluding the SipA protein which is specifically required for ascospore formation in fruit bodies in A. nidulans. Therefore, levels of the fungal mycotoxin sterigmatocystin (ST) were measured in mutants by HPLC (Fig 4D). Production of ST was significantly reduced in mutants in comparison to the WT. strA and sipA mutants produced less ST than WT but more than sipB, sipC, sipD and sipE (ppgA) mutants. ST production of the strA/sipA double mutant was slightly less than the respective single mutants. However, double deletion of strA with sipB, sipC, or sipE resulted in extremely reduced ST (less than 10% of WT).
In order to examine if the drastic alterations in SM production in AnSTRIPAK mutants were due to changes in transcript levels of the SM regulators, the expression profiles of the GMM plates and incubated for 5 days at 37˚C under constant light (90 μWm 2 ) or dark conditions. Upper squares show colony development (scale 1 cm) from plates and lower squares are close-up images of the colonies (scale 100 μm). Whitish round structures represent fruit bodies. sAΔ represents strAΔ, AΔ represents sipAΔ, BΔ represents sipBΔ, CΔ; sipCΔ, DΔ; sipDΔ, EΔ; sipEΔ. For consistency, ppgAΔ was presented as sipEΔ (B) Quantification of comparative radial growth rate of the single and double mutants from light induced plates. NS; not significant (p > 0.05), S; significant (p < 0.05). WT growth rate serves as 100% standard. (C) Quantification of asexual and sexual development from single and double STRIPAK mutants. L; light, D; Dark, S; significant. Asexual sporulation of WT in light represents 100% sporulation. Sexual fruit body formation of WT in dark was used as 100%. Values are the means of three replicates, and error bars represent standard errors.
https://doi.org/10.1371/journal.pgen.1008053.g002 velvet complex along with ST gene cluster as well as two additional SM clusters were determined. Expression of the velvet complex (veA/velB/laeA) was generally diminished in the absence of AnSTRIPAK complex (Fig 4E). The most drastic decrease was observed in velB and veA expression. The reduced expression of the velvet complex is translated into expression of the ST gene cluster since expression of the transcription factor aflR and the two structural genes stcQ and stcE sharply dropped in AnSTRIPAK mutants except for sipAΔ ( Fig 4F). Antibiotic penicillin (PN) and antitumour terrequinone (TQ) genes were tested in addition. Surprisingly, expression of acvA, ipnA and aatA required for PN production was reduced in AnSTRIPAK mutants. ipnA expression in sipBΔ and aatA expression in sipDΔ did not change ( Fig 4G). Similarly, tdiA and tdiB genes of the TQ cluster were significantly down-regulated in AnSTRIPAK mutants except for the sipA mutant, which was higher ( Fig 5E). These metabolite and expression data reveal that the AnSTRIPAK complex is important for production of ST and expression of the velvet complex. Furthermore, full expression of at least three different gene clusters ST, PN and TQ require an intact AnSTRIPAK complex, except for sipA.

Localization of the AnSTRIPAK complex on the nuclear envelope is dependent on striatin
Striatin (StrA) is localized to the nuclear periphery and endomembrane systems in A. nidulans [26] (S1 Fig). However, it is not known if the entire STRIPAK complex is also associated with the nuclear envelope in A. nidulans. We have found that SipA to SipF interact with StrA constituting the heptameric AnSTRIPAK complex. A functional StrA-mRFP fusion was coexpressed with SipA, SipB, SipC, SipD and SipE-GFP fusions ( Fig 5). The StrA-mRFP fusion, which clearly decorated the nuclear periphery where the nuclear envelope is found, was not primarily found inside the nucleus. Furthermore, StrA was also present on long string-like extensions, presumably representing endomembrane systems such as endoplasmic reticulum. The SipA-GFP fusion was also found to be accumulated around the nucleus and in string-like extensions but was present in the nucleus at trace levels and overlapped with StrA-mRFP signals ( Fig 5A). Similarly, GFP fusions of SipB and SipC were found to be co-localised with StrA around the nucleus (Fig 5C). In co-localizations with histone H2A, SipD-GFP and SipE-GFP showed clear perinuclear localization, very similar to those of SipA-GFP, SipB-GFP and SipC-GFP (Fig 6). Colocalizations of SipD and SipE (PpgA)-GFP with StrA-mRFP showed a somewhat different pattern. The staining was less punctate and the localization around the nucleus was visible but weaker than that of SipA-to SipC-GFP (Fig 5D and 5E). We speculate that the mRFP tag on StrA weakens the binding of SipD and SipE to StrA.
Like StrA, SipA to SipE were all co-localized with StrA around the nuclear envelope and partially in endomembrane systems. Since sip double deletions with strA led to more sensitive phenotypes suggesting the key role of StrA for the molecular function of the AnSTRIPAK complex, SipA to SipE-GFP fusions were expressed in a strain devoid of StrA (Fig 6). The absence of StrA did not influence expression of the Sip proteins except for SipD, which showed a higher molecular weight as well as a thicker lower molecular band ( Fig 6D). Interestingly, lack of StrA led to loss of nuclear envelope localization of SipA, which became more diffuse in com.sipCΔ, com.sipDΔ, com.sipEΔ) in the presence of hydrogen peroxide (H 2 O 2 , 1 mM) measured as radial growth rate. (B) Oxidative stress responses of the strains to Menadione (0.08 mM). (C) Growth of the mutants in the presence of cell-wall stress agent Congo red (20 μg/ml). The cultures (5x10 3 spores) were grown on solid GMM with stress agents for 5 days at 37˚C in light. Scale bar 1 cm. These experiments were repeated at least three times with the same results. Chart graphs show radial growth diameter compared with the WT, which was used as standard (100%). Data are indicated as average ± SD of three independent biological repetitions. Columns with (ns) denote non-significant but (n) denote significant difference and also ( ��� ) represent values of strong significant difference p < 0.0001 compared with WT.
https://doi.org/10.1371/journal.pgen.1008053.g003  the cytoplasm. SipB-GFP also lost its nuclear envelope localization in the absence of StrA and dispersed in the cytoplasm and, interestingly, was present in the nucleoplasm except for the nucleolus (Fig 6B). SipC also dispersed from the nuclear envelope and was relatively uniformly distributed in the cytoplasm but it was at least partially excluded from nuclei ( Fig 6C). SipD remained punctate in the absence of StrA, but it was no longer concentrated at the nuclear envelope, and many punctae were seen in the nucleoplasm ( Fig 6D). SipE (PpgA) showed a similar localization pattern to SipB in the absence of StrA, diffuse in the cytoplasm and nucleoplasm but excluded from the nucleolus (Fig 6E). In summary, our expression data and confocal imaging data reveal that all of the members of the AnSTRIPAK complex localize to the nuclear envelope and endomembrane system. They require the molecular scaffold StrA for normal localization and all but SipD disperse in the absence of StrA. SipD does not disperse in the absence of StrA but its localization pattern is altered.

Lack of Striatin disrupts the assembly of the AnSTRIPAK complex and reveals subcomplex dynamics
Given the fact that StrA is required for appropriate cellular localization of the AnSTRIPAK complex, it became intriguing to ask whether the mislocalizations reflect the interdependent interactions of the complex proteins. As assayed by TAP purification followed by MS, SipA to SipE all complexed with each other with high peptide numbers in the presence of StrA (Fig 6F, S13-S17 Tables). However, surprisingly, in the absence of StrA, TAP purification of SipA did not pull down any other members of the complex. SipB and SipD only reciprocally copurified with each other. SipC and SipE (PpgA) reciprocally pulled down the regulatory subunit of phosphatase SipF as well as a second regulatory subunit B (PabA; presumably reflecting a distinct, STRIPAK-independent PP2A complex). Surprisingly, in the absence of SipA, StrA was able to establish a form of the AnSTRIPAK complex lacking only SipA (S18 Table). These TAP data comparing the physical interaction dynamics of AnSTRIPAK complex in the presence and absence of StrA clearly display that (I) SipA only interacts with StrA, therefore it is recruited to the AnSTRIPAK complex via StrA and StrA does not need SipA to establish the AnSTRIPAK complex, (II) SipB-SipD form heterodimers and then presumably associate with StrA, (III) SipC-SipE-SipF form a heterotrimeric subcomplex, which is then attached to StrA to form the fully functional AnSTRIPAK complex.

Striatin controls two unrelated MAPK-pathways: nuclear accumulation of the stress-sensing MpkC MAP kinase and activation of the developmentinducing MpkB MAP kinase
It was shown that activation of the N. crassa cell-wall stress pathway MAPK MAK-1 as well as its nuclear accumulation was reduced in core components of STRIPAK complex mutants [15]. The same study also showed that MAK-2, which regulate hyphal fusions in N. crassa, phosphorylates MOB-3 component of STRIPAK. Striatin mutants in A. nidulans show developmental and SM defects. Furthermore, the mutants also show sensitivity to cell-wall and oxidative stressors. Therefore, we wondered how common MAPK pathways are influenced by the STRIPAK complex and whether the influence of STRIPAK on MAPK pathways was similar to that of N. crassa. In addition to MpkB which controls sexual development and SM production, there are three mitogen activated protein kinases (MAPK): MpkA is mainly responsible for cell-wall regulation, MpkC and SakA (yeast Hog1 ortholog) play roles in stress responses, particularly oxidative and osmotic stress responses [24,28]. To determine if StrA influences the localizations of these kinases, the three MAPKs, MpkA (cell-wall stress), MpkB (sexual development and SM production) and MpkC (oxidative and osmotic stress) were expressed as GFP fusions in the presence and absence of StrA (Fig 7). All kinases were expressed similarly in the presence or absence of StrA ( Fig 7A). Surprisingly, activation phosphorylation (P-44/42) of MpkB, which is necessary for fruit body formation, was almost totally lost in the absence of StrA. MpkB localization, which was not influenced by lack of StrA, was mainly nucleo-cytoplasmic ( Fig 7B). Furthermore, MpkB (yeast Fus3p) interacted with MAP2K MkkB (yeast Ste7p), recently characterized scaffold protein HamE and transcription factor SteA (yeast Ste12p) in the absence of StrA (S19 and S20 Tables). In contrast to N. crassa MAK-1, MpkA exhibited nuclear localization both in WT and strAΔ strains, which was not influenced by cell-wall damaging antifungal drug caspofungin ( Fig 7C). Interestingly, MpkC, which was mainly found in the cytoplasm under non-stress conditions in the WT, imported into the nucleus in the absence of StrA (Fig 7D). Caspofungin treatment had no influence on the localization of MpkC. Both oxidative and osmotic stress conditions led to nuclear accumulation of MpkC in WT. However, lack of StrA resulted in loss of MpkC nuclear enrichment under oxidative stress whereas osmotic stress had no effect on MpkC localization in the strA deletion. These results imply that StrA acts differently in A. nidulans and mainly prevents MpkC from entering into the nucleus under non-stress conditions to respond to stress and is required for nuclear accumulation of MpkC under oxidative stress conditions. Furthermore, StrA controls phosphorylation of MpkB, which is required for coordination of development and SM production.

Discussion
The STRIPAK complex is highly conserved in eukaryotes and is involved in many cellular functions [1,6]. In this study, we have revealed the molecular nature and functions of the AnSTRIPAK complex, which consist of at least seven proteins StrA/STRN, SipA/Phocein, SipB/SIKE1-like, SipC/STRIP, SipD/SLMAP, SipE(PpgA)/PP2Ac and SipF/PP2AA. Detailed phenotypic, genetic, biochemical, live cell imaging and chemical approaches identified StrA as the core scaffolding protein which assembles at least six other members at the nuclear envelope to control intracellular signaling events (Fig 7F). StrA is found in a heterodimer state with SipA/Phocein and further recruits the heterodimer SipB/SipD (SIKE/SLMAP) and the heterotrimer SipC/SipE/SipF (STRIP/PP2Ac/PP2AA) to establish a heptameric complex. This complex controls gene expression for the sexual reproductive cycle, including formation of multicellular fruit bodies and furthermore, it is also a key regulator for production of SMs.
Although the molecular composition of the STRIPAK complex is conserved, the described functions of the complex show diverse roles in fungi, flies and mammals. In the baker´s yeast S. cerevisiae, the FAR (STRIPAK) complex acts as an antagonist of target of rapamycin (TOR) pathway and counteracts recovery from pheromone arrest [11,12] whereas in the fission yeast S. pombe, the STRIPAK complex acts as a negative regulator of septation initiation [13]. In closely related filamentous fungi S. macrospora and N. crassa, STRIPAK regulates hyphal fusion and fruit body formation [14,15]. In two Fusarium species, which are plant pathogenic fungi, striatin is required for pathogenicity on host organisms [29,30]. In the fruit fly, the STRIPAK complex controls epithelial cell movement and tissue size by modulating two different signaling pathways [31,32], whereas in the nematode C. elegans, members of the complex control polarity establishment during embryogenesis [33]. In humans, STRIPAK complex governs embryonic stem cell differentiation, proper cardiac function, dendritic spine morphology and cancer [34,35].
Lack of fully assembled AnSTRIPAK complex results in loss of proper light response as a result of defective asexual and sexual development. The AnSTRIPAK complex controls lightdependent fungal development. In this fungus, light controls asexual reproduction through the various light receptors. The velvet complex physically and functionally interacts with the red and blue light receptors [36]. Improper expression of the velvet complex in STRIPAK mutants might influence interaction dynamics of the complex with light receptors, therefore, disrupting the light-dependent development. Particularly, two major asexual transcription factors abaA and brlA are induced by AnSTRIPAK complex, which drives asexual responses. The role of AnSTRIPAK complex in sexual development might be somewhat complicated. It controls formation of fruit bodies by properly dosing expression of sexual transcription factors such as nsdD and steA. It is known from N. crassa and S. macrospora that the STRIPAK complex is involved in cellular fusion, which finally leads to fruit body formation in these fungi. In A. nidulans, fruit body formation is also controlled by formation of cell-cell fusions. Loss of fruit bodies also indicate that there are defects in cell-cell fusions in the absence of AnSTRIPAK. In N. crassa, the NRC-1-MEK-2-MAK-2 kinase cascade are the central components of self signalling machinery [37]. The MOB-3 component of the NcSTRIPAK complex interacts with MAK-2. However, nuclear localization of MAK-2 is not influenced by STRIPAK but MAK-1 localization is altered. In A. nidulans, the sexual pathway is controlled by a pheromone response (SteC-SteD-MkkB-MpkB) module which migrates from the plasma membrane to the nuclear envelope to deliver MpkB into the nucleus. Phenotypes of MAPK module mutants are very similar to STRIPAK mutants [24]. Interestingly, in time course purifications at sexual stages, SteD, which is the adaptor domain of the pheromone response pathway, was repeatedly co-purified with the StrA-TAP fusion (S1-S6 Tables). Both MkkB and MpkB localizations did not change and MpkB interacted with both MkkB and SteA in the absence of StrA. However, interestingly, MpkB was not activated by phosphorylation in the absence of StrA. Since MpkB interacts with SteA-VeA and phosphorylates VeA, which is necessary for sexual development Fungal STRIPAK complex assembly and SM production, the AnSTRIPAK complex presumably activates fruit body formation and SM production by phosphoactivation of the MAPK module.
The sipA mutant and its strA combinations showed an opposite phenotype to STRIPAK component deletions, indicating an epistatic function of SipA over the StrA but not over the SipB-SipC-SipD-SipE complex. The sipA deletant grows as well as WT or even better. It produces 2.5-fold more asexual spores and is more resistant to different stress conditions. SM production was slightly influenced in the absence of sipA and expression of all other gene clusters was upregulated in sipAΔ, which strongly suggests an inhibitor role for SipA in these processes. How does SipA perform this function within the AnSTRIPAK complex? This is obvious from interaction dynamics, because SipA interacts only with StrA during all developmental stages. Furthermore, in the absence of SipA, StrA is able to form a hexameric complex with SipB, SipC, SipD, SipE and SipF. This partial AnSTRIPAK(-SipA) complex without SipA is presumably more active than an intact complex, and it promotes excessive growth and asexual sporulation by unknown mechanisms. However, it does not sufficiently fulfill the meiotic functions of the WT complex, because the sipA deletant cannot produce ascospores. MOB-3 (SipA homolog) in N. crassa interacts with the NRC-1-MEK-2-MAK-2 kinase self-signaling cascade [38]. However, although SipA does not interact with the components of the MAPK pathway in the absence of StrA in TAP studies, it might transiently interact with the MAPK pathway to elicit its effect on development. The partial complex functions (AnSTRIPAK(-SipA)) will require more understanding at experimental level.
In N. crassa and S. macrospora, the involvement of the STRIPAK complex in SM has not been reported. Production of the mycotoxin ST is positively controlled by the AnSTRIPAK complex, which requires proper expression of the velvet complex. Accordingly, the expression of the ST gene cluster is also drastically reduced by loss of the AnSTRIPAK complex. Expression of PN and TQ clusters are similarly diminished in the STRIPAK mutants. A. nidulans produces many more metabolites than these three molecules. Although only three gene clusters were examined here, the effects of the loss of STRIPAK might be more extensive. How does the complex control SM production? Transcriptional downregulation of the velvet complex might be the primary reason why SM production is drastically affected in STRIPAK mutants. Another scenario might be that AnSTRIPAK is important for vital functions of the SteC-SteD-MkkB-MpkB module. Because this module uses the nuclear envelope to interact with the nucleus and deliver the active MAPK MpkB into the nucleus. As discussed previously, reduction in signal fidelity of the pheromone response pathway in the absence of StrA presumably results in reduced SM production.
In other eukaryotes, STRIPAK complex acts as a negative regulator of kinases, because the GCKIII kinase family member Mst3 and Mst4 are hyperphosphorylated in the mutants of PP2A subunit in human cells or okadaic acid treated cells, respectively [39]. The scenario in A. nidulans and N. crassa is different, however. In A. nidulans MpkB loses its phosphorylation in the absence of striatin and MpkB is a MAPK, not a GCK type kinase. In contrast, in N. crassa, MAK-1 loses its activity under resting and stress conditions [15]. Two GCKs SmKIN3 and SmKIN24 were found to be functionally and physically interacting with S. macrospora striatin [40,41]. However, functional control of these kinases by striatin remains to be shown.
The SIKE-like domain containing SipB is part of the AnSTRIPAK complex and it forms a heterodimer with SipD (SLMAP), which is similar to mammals where SIKE and SLMAP form a heterodimer. However, it is not surprising to see a mammalian anti-inflammatory protein conserved in fungi, including A. nidulans since it was also shown that velvet family proteins of A. nidulans contain a DNA binding domain structurally similar to the proinflammatory Rel family of NF-KB proteins although they show only 13% similarity at the amino acid level [42]. Presence of NF-KB like velvet family proteins as well as SIKE-like proteins in fungi strengthens the hypothesis that fungal and mammalian defense systems have some degree of analogy since SMs of fungi often act as a defense mechanism against competitors or predators [43]. A SIKElike protein as part of the AnSTRIPAK complex participates in production of SMs such as the mycotoxin ST to help the fungus to compete with other organisms in soil.
The AnSTRIPAK complex mutants are sensitive to oxidative and cell-wall stress, however they are not extremely influenced by osmotic stress. At least two genes catC and sodB are down-regulated in the absence of the AnSTRIPAK complex. In N. crassa, transport of MAK-1 (cell wall regulator kinase) into the nucleus is facilitated by NcSTRIPAK complex in a MAK-2-dependent manner. However, the scenario is somewhat different in A. nidulans. A. nidulans stress responses are mediated by three MAPK proteins, MpkA, SakA and MpkC. MpkA regulates cell-wall integrity [28]. SakA and MpkC MAPKs control oxidative stress responses interdependently in A. nidulans [44,45]. Striatin has no influence on MpkA, which is a homologue of N. crassa MAK-1. MpkC nuclear accumulation is restricted by StrA in the absence of stress conditions. It is known that MpkC orthologs are activated in the cytoplasm in response to stress and enter into the nucleus. The function of the AnSTRIPAK complex is presumably to keep MpkC in the cytoplasm in the absence of stress, which allows subsequent activation under stress conditions.
In conclusion, this study revealed the composition and assembly hierarchy of the AnSTRI-PAK complex. It was surprising that STRIPAK also regulates light and stress responses and the production of SMs. Since A. nidulans is representative of more than three hundred biotechnologically and medically relevant members of the genus Aspergillus, the information gained through this study is applicable to other important fungi to understand their biology and to use their full potential in SM and enzyme production. Paradoxically, deletion of sipA resulted in an increase in transcription of a number of SM genes and this deletion may be a useful tool in eliciting expression of silent gene clusters, a goal of genome mining for useful SMs. A functional link between STRIPAK and MAPK pathways has been established and further information on this connection will also reveal how fungi sense various signals and how they control signal influx on the nuclear envelope and how they convert these signals into appropriate responses to control their growth, development and pathogenicity.

Strains, media and growth conditions
Fungal strains used in this study are listed in S21 Table. A. nidulans AGB551, which has a WT veA allele, was used for all deletion and epitope taggings. Stellar (Clontech) and MACH-1 (Invitrogen) competent Escherichia coli cells were used for recombinant DNA preparations. The WT and transformed A. nidulans strains were grown in Glucose Minimal Medium (GMM), supplemented with appropriate amounts of vitamins. For vegetative stage experiments, fungal spores were grown submerged in liquid GMM with 180 rpm rotation for 20 or 48h. For induction of development, vegetative mycelia were then filtered through miracloth and placed on solid GMM with 2% Agar. To induce the cultures asexually, cultures were grown vegetatively for 20h and shifted to the plates and were further incubated in the presence of light for 6, 12, 24h. For sexual induction, plates were covered by aluminium foil and incubated in the dark for 6, 12, 24, 48h. E. coli were grown in LB broth agar or liquid LB with Ampicillin (100 μg/ml) at 37˚C overnight.

Nucleic acid and plasmid constructions methods
Transformation of E. coli and A. nidulans were performed as explained in detail [46,47]. The plasmids and oligonucleotide sequences used and created in this study are listed in S22 and S23 Tables, respectively.
To create pNE17, pNE18 and pNE25, the sipC promoter and ORF was amplified with NE15/65 and the terminator region with NE19/66, using AGB551 genomic DNA as a template. These fragments were fused to sgfp::AfpyroA, sgfp::AfpyrG or ctap:: AfpyrG cassettes and cloned in pUC19 as described for the sipA and sipB. The primers NE16/20 amplified 10 kbp linear fragments from (pNE17, pNE18 and pNE25). These fragments were transformed into the wild type to generate ANNE18, NANE23, NANE28, respectively. Gene replacement events were verified by the Southern hybridization.

Hybridization techniques
Southern hybridization experiments were performed as given in detail [48]. Fungal genomic DNA was prepared from plates using the ZR Fungal/Bacterial DNA MiniPrep TM (Zymo Research) kit. 700 ng of isolated genomic DNA was used for restriction enzyme digestion. The southern hybridization was performed with non-radioactive probes by using DIG labeling (Roche) as described in the user protocol.

Phenotypic assays
The phenotypes and quantification of sporulation for WT and deletion strains were examined as follows: Fungal spores were counted using a haemocytometer. 5 x10 3 spores (5μl) were used to point inoculate solid GMM, containing appropriate supplements. Plates were incubated in the light (for asexual development) and in the dark (for sexual development) for 4-5 days at 37˚C. Colonies were observed using the Olympus szx16 microscope with Olympus sc30 camera. Digital pictures were taken and processed with the Cell Sens Standard software (Olympus). Quantifications were performed from three independent biological replicates.

Protein extraction
Fungal mycelia were obtained from liquid cultures and broken using liquid nitrogen. Protein extracts were prepared by re-suspending the smooth mycelia in protein extraction buffer (B300 buffer) which contains 300 mM NaCl, 100 mM Tris-Cl pH: 7.5, 10% Glycerol, 1 mM EDTA, 0.1% NP-40 and added with 1mM DTT, 1x Protease inhibitor mix (Roche), 1.5mM Benzamidine, 1x Phosphatase inhibitor mix (Roche) and 1mM PMSF. Protein concentrations were calculated by performing Bradford assays. 100 μg of total protein extract was run on various percentages SDS gels as required (10% or 12%) and transferred to protean membrane with 0.45μm pore size (GE Healthcare).

Tandem Affinity Purification (TAP), GFP pull-downs and LC-MS/MS Protein identification
TAP experiments, GFP pull-downs and preparation of the protein crude extracts and analysis of the proteins were performed as explained in detail [23,49]. TAP experiments were performed from two independent biological replicates. TAP of the WT strain eluates were used as non-specific control. Proteins identified from StrA, SipA, SipB, SipC, SipD and SipE TAP experiments were filtered from non-specific control purifications. The final lists of the proteins were presented in excel format in Supplementary tables (S1-S6 and S8-S18 Tables). GFP pulldowns of MpkB interaction partners in the presence or absence of StrA originates from either one or two biological replicates (S19 and S20 Tables).

Confocal microscopy
Green and monomeric red fluorescent protein (GFP and mRFP) expressing strains were grown in 500 μl liquid GMM media, with appropriate supplements in sterile Lab-Tek Chambered Coverglasses with covers (8 wells per coverglass) (Thermo Scientific) for 17-20 hours at 30˚C. Localizations of the proteins were captured and recorded as published previously [51].

RP-HPLC analysis of secondary metabolites
The WT and all deletion strains were grown as follows: 5x10 3 spores were inoculated on GMM supplemented with vitamins and 1% Oatmeal at 37˚C for 5 days. Metabolites were extracted by using chloroform and drying them in speed vac. Samples were suspended in 200μl Methanol. 1mg/ml sterigmatocystin (Sigma) was used as the standard (2.5 μl of standard added in 47.5 μl of 100% Methanol). RP-HPLC analysis was carried out on a Shimadzu RP-HPLC with a photodiode array detector (PDA). 20 μl of standard or sample was injected onto a Luna omega 5μm polar C18 (LC column 150 x 4.6 mm) and separated across a water: acetonitrile gradient with 0.1% (v/v) TriFluoroacetic Acid (TFA). Gradient conditions of 5-100% acetonitrile over 30 min with a flow rate of 1 ml/min were used with PDA detection at 254 nm.

Quantitative real time PCR
100 mg of mycelia was used for RNA isolation by using RNeasy (Qiagen). 1 μg RNA was used for each cDNA synthesis using the Transcriptor First Strand cDNA Synthesis Kit (Roche). Different primers of target genes were used for qPCR reaction as described in the user protocol, by using LightCycler 480 Sybr Green I Master (Roche). A house-keeping gene, benA was used as a standard. Relative Expression Analysis was performed by Light Cycler 480 Software.

Statistical analysis
All experiments were performed on three independent occasions and numerical data (S1 Dataset) are expressed as the mean ± SD and standard error. The corresponding means were compared for significant differences via the student t-test and One-way ANOVA methods, by using the software Graphpad Prism Version 6. The cultures (5x10 3 spores) were grown for 5 days at 37˚C in light. Scale bar 1 cm. These experiments were repeated at least three times. Chart graphs show radial growth diameter compared with WT, which was used as standard (100%). Data are indicated as average ± SD of three independent biological repetitions. Columns with (ns) denote non-significant but (n) denote significant difference and also ( ��� ) represent values for strong significant difference (P<0.0001) compared with WT. (TIF)