Germinal Center Kinases SmKIN3 and SmKIN24 Are Associated with the Sordaria macrospora Striatin-Interacting Phosphatase and Kinase (STRIPAK) Complex

The striatin-interacting phosphatase and kinase (STRIPAK) complex is composed of striatin, protein phosphatase PP2A and protein kinases that regulate development in animals and fungi. In the filamentous ascomycete Sordaria macrospora, it is required for fruiting-body development and cell fusion. Here, we report on the presence and function of STRIPAK-associated kinases in ascomycetes. Using the mammalian germinal center kinases (GCKs) MST4, STK24, STK25 and MINK1 as query, we identified the two putative homologs SmKIN3 and SmKIN24 in S. macrospora. A BLASTP search revealed that both kinases are conserved among filamentous ascomycetes. The physical interaction of the striatin homolog PRO11 with SmKIN3 and SmKIN24 were verified by yeast two-hybrid (Y2H) interaction studies and for SmKIN3 by co-Immunoprecipitation (co-IP). In vivo localization found that both kinases were present at the septa and deletion of both Smkin3 and Smkin24 led to abnormal septum distribution. While deletion of Smkin3 caused larger distances between adjacent septa and increased aerial hyphae, deletion of Smkin24 led to closer spacing of septa and to sterility. Although phenotypically distinct, both kinases appear to function independently because the double-knockout strain ΔSmkin3/ΔSmkin24 displayed the combined phenotypes of each single-deletion strain.


Introduction
In filamentous fungi, the development of multicellular fruiting bodies requires highly conserved differentiation processes and is essential for sexual reproduction [1]. The filamentous ascomycete Sordaria macrospora is a well-established model organism for studying this process [2,3]. S. macrospora is a haplont with a homothallic life style and does not require a mating partner for sexual reproduction [4]. In contrast to many other filamentous ascomycetes, S. macrospora produces no asexual spores [5]. The life cycle begins with the germination of  [42]. Based on structural similarity to the S. cerevisiae sterile-20 protein kinase (Ste20), GCKs are members of the Ste20-related group of protein kinases. These are sub-classified into p21-activated kinases (PAK) and GCK families based on the location of their kinase domain [43], which is C-terminal or N-terminal to the regulatory domain, in PAKs and GCKs, respectively [44]. GCKs share a highly conserved catalytic domain and a more divergent regulatory C-terminus [43,44]. The GCK family can be further subdivided into subfamilies GCK I to GCK VIII [44].
In mammals, GCKs III MST4, STK24, STK25 and the GCK IV MINK1 are identified as components of the STRIPAK complex [15][16][17]45]. CCM3 has recently been shown to act as an adaptor that recruits kinases of the GCK III family to the striatins [15,17,45]. The role of CCM3 is to bring GCK III into close proximity to striatin-connected phosphatase PP2A whereas GCK IV MINK1 is directly associated with striatin [16,45]. It is thought that the PP2A phosphatase associated with the STRIPAK complex dephosphorylates the GCKs, and thereby reduces their catalytic activity [16,45].
In order to identify STRIPAK kinases in S. macrospora, we performed a BLASTP search with the mammalian STRIPAK kinases MST4, STK24, STK25 and MINK1 against the S. macrospora genome. We identified two GCKs, SmKIN3 and SmKIN24, as possible homologs of the mammalian STRIPAK-associated GCKs. Y2H analyses revealed that both kinases could interact with the S. macrospora striatin homolog PRO11 and the interaction between SmKIN3 and PRO11 was verified in vivo by co-IP. Fluorescence microscopy revealed that both proteins localize to septa. Deletion of Smkin3 led to fewer septa and increased protoplast regeneration and aerial hyphae formation, whereas deletion of Smkin24 caused hyperseptation and sterility. Despite their distinct phenotypes, both kinases appear to act independently because the ΔSmkin3/ΔSmkin24 double-knockout exhibited a phenotype that combined those of the single-deletion strains.

Strains, media and culture conditions
All strains used in this study are listed in Table 1. Chemically competent Escherichia coli Mach1 was transformed according to a standard protocol [46]. Plasmids (Table A in S1 File) were generated by homologous recombination in yeast strain PJ69-4A [47] as described in [48] or with the In-Fusion HD Cloning Kit (Clontech, 639648). Transformation of yeast cells was performed by using an Eppendorf Electroporator 2510 (Eppendorf, Hamburg, Germany) at 1.5 kV [49]. S. macrospora transformation was done as described previously [50,51] and selected on media supplemented with nourseothricin-dihydrogen sulfate (50 μg/ml) (WernerBioAgents, 5004000) or hygromycin B (110 U/ml) (Merck, 400051-10MU). S. macrospora strains (Tab 1) were grown in liquid or on solid corn meal (BMM) or fructification medium (Sordaria Westergaard) SWG [52,53] at 27°C. Growth velocity of the respective strains was measured in 30 cm race tubes filled with 25 ml solid SWG medium. The average growth velocity represents the results of three independent experiments, each performed in triplicate.
Phusion High-Fidelity DNA polymerase (New England Biolabs, M0530S) were used according to the manufacturer's protocol. All primers listed in Table B in S1 File were purchased at Eurofins MWG Operon (Ebersberg, Germany).

RT-PCR
RNA from S. macrospora was isolated according to [55]. cDNA was generated via reverse transcription reaction using Transcriptor High Fidelity cDNA Synthesis Kit (Roche, 05081955001). Template concentration was adjusted to 2 μg RNA per reaction.

Generation of S. macrospora knockout strains
Deletion strains were generated by transforming the respective deletion cassette into S. macrospora strain Δku70 which is impaired in non-homologous end joining [54]. For the construction of the deletion cassette, the 5' and 3' flanking regions of the respective genes were amplified via PCR with primers (Table B in S1 File) having 29-bp overhangs homologous to the hygromycin cassette as well as to the pRSnat or pRShyg plasmid backbone [31,56]. The hygromycin resistance cassette hph was amplified from plasmid pCB1003 [57]. All fragments were merged and inserted into plasmid pRSnat or pRShyg by homologous recombination in the S. cerevisiae strain PJ69-4A [48]. This resulted in a deletion construct comprising the 5' region, followed by hph and the 3' region of the gene that should be deleted. The respective deletion constructs were entirely amplified via PCR, purified and transformed into S. macrospora Δku70. To eliminate the Δku70 (nourseothricin resistance) background S. macrospora primary transformants were crossed with spore-color mutant r2 [3]. Afterwards, hygromycin resistant, nourseothricin sensitive single-spore isolates were selected. Double-deletion strain ΔSmkin3/ΔSmkin24 was generated by crossing the single deletion strains ΔSmkin3 and ΔSmkin24. Deletion strains derived from single-spore isolates were verified by PCR using primers enlisted in Table B in S1 File and Southern hybridization using AlkPhos Direct Labelling and Detection Kit (GE Healthcare, RPN3690) [46].

Co-immunoprecipitation
The plasmid encoding for the N-terminally HA-tagged PRO11 under control of the constitutive N. crassa ccg1 promoter was described previously [31]. pFLAG-SmKIN3 encoding the Nterminally FLAG-tagged SmKIN3 was constructed by combining the full length cDNA of Smkin3 (see version I of Smkin3 transcripts Fig A in S1 File), with the N. crassa ccg1 promoter of flowed by 3xFLAG [60,61] and the trpC terminator of Aspergillus nidulans into vector pRShyg [31]. Plasmids encoding for HA-PRO11 [31] and FLAG-SmKIN3 were separately transformed into S. macrospora wt strain and combined by crossing resulting in a hygromycin and nourseothricin resistant strain expressing genes coding for both constructs. Protein purification was done as described above. The samples were separated by SDS-PAGE and blotted onto nitrocellulose membrane. For immunodetection, HA and FLAG antibody and HPR substrate were used as described above. The experiment was repeated two times independently. Co-IP was conducted with additional use of the thiol crosslinking reagent Bismaleimidohexane (BMH) (Thermo scientific, 22330) solved in Dimethylsulfoxide (DMSO = at a concentration of 20 mM. The crosslinker was used in a final concentration of 0.2 mM. The reaction was stopped after incubation for 2 h at 4°C by adding Dithiothreitol (DTT) in a final concentration of 25 mM as recommended in the manufacturers protocol.
Yeast two-hybrid analysis Y2H analysis in S. cerevisiae was conducted with the Matchmaker two-hybrid system 3 (Clontech, USA). pBD-SmKIN3 and pBD-SmKIN24 were generated by using In-Fusion HD Cloning Kit (Clontech, 639648). Smkin3 (version with all introns spliced, version I in Fig A in S1 File) was amplified from cDNA with primers kin3_pBD_inf_F and kin3_pBD_inf_R, Smkin24 (transcript version with all introns spliced, version I in Fig A in S1 File) was amplified from cDNA with primers kin24_pBD_inf_F and kin24_pBD_inf_R and inserted into vector pGBKT7. Plasmid pBD-SmKIN3 and pBD-SmKIN24 were transformed into yeast strain Y187, pAD11FL containing full-length pro11 was transformed into AH109, respectively (Tab 1).
The respective Y187 and AH109 strains were mated as previously described (45) and selected on solid SD-medium lacking leucine (leu) and tryptophan (trp). Y187 and AH109 strains are adenine (ade), histidine (his), leu and trp auxotroph (Tab 1). Leu auxotrophy is restored with pGADT7 plasmid and derivatives, trp auxotrophy with pGBKT7 and derivatives. In AH109 the reporter gene ade2 is under control of the GAL4 regulated GAL2 promoter, but the gene coding for the GAL4 transcription factor is deleted (Tab 1). Proteins tested for interaction are fused either to GAL4 binding domain (BD) (vector pGBKT7) or GAL4 activation domain (AD) (vector pGADT7). Interaction of these proteins leads to close proximity of GAL4 DNA binding and activation domain and enable expression of the reporter gene. Growth of mated Y187-AH107 diploids on medium lacking leu, trp and ade, requires presence of both plasmids and interaction of the respective GAL4 fusion proteins. Positive colonies from SD-medium without leu and trp were used for drop plate assays. A serial dilution of cells was spread on SD agar plates without leu, trp and ade. To prove whether the respective genes in vector pGBKT7 were expressed properly, a test based on RanBPM was used by mating Y187 transformants with AH109 carrying pAD-RanBPM [62]. RanBPM interacts with the GAL4-binding domain and thus functions as expression control of the gene fused to GAL4 binding domain (54). pBD-SmKIN3 and pBD-SmKIN24 were also checked for self-activation as described in [63].

Light and fluorescence microscopy investigations
To localize SmKIN3 and SmKIN24 in S. macrospora, N-terminally eGFP-tagged constructs under control of the A. nidulans gpd promoter were generated and introduced into their respective deletion strain. Since transcripts of Smkin3 and Smkin4 were alternatively spliced, plasmids encoding for SmKIN3-eGFP and SmKIN24-eGFP were constructed by amplifying the ORF of Smkin3 or Smkin24 corresponding to the version with all spliced introns of the transcript (version I in Fig A in S1 File). The cDNA was inserted into NotI digested pDS23-egfp vector [64] via homologous recombination in yeast. The respective primers, used for amplification of Smkin3 (Smkin3ngfp_F x Smkin3ngfp_R) and Smkin24 (Smkin24ngfp_F x Smkin24ngfp_R) are listed in Table B in S1 File. Hyphae, sexual structures and fluorescence were visualized by an AxioImager M1 microscope (Zeiss, distributed by Visitron Systems GmbH) combined with a Photometrics CoolSNAP 2 HQ camera (Roper Scientific, Photometrics). The respective pictures were processed with Metamorph (version 6.3.1; Universal Imaging), GIMP 2.8.2. (GNU Image Manipulation Program, The GIMP Development Team) and Illustrator CS2 (Adobe). To display eGFP or Calcofluor white fluorescence chroma filter set 49002 and 49000 were used, respectively. Calcofluor white staining for visualization of cell walls and septa was conducted by adding 40 μl Calcofluor white (Sigma-Aldrich, 18909), freshly diluted 1:1 with 10% KOH solution and directly applied on the mycelium.

Phylogenetic analysis
Multiple protein sequence alignments were performed using the ClustalX2 program [65]. Phylogenetic analysis was made with programs from package PHYLIP version 3.695 (http:// evolution.genetics.washington.edu/phylip.html). Distance matrices were calculated using the program PRODIST and were then used for constructing phylogenetic trees with the neighborjoining (NJ) program NEIGHBOR. To evaluate the statistical significance a bootstrap analysis with 1000 iterations of bootstrap samplings and reconstruction of trees was performed. A majority rule consensus tree was subsequently generated with the program CONSENSE, displayed with TreeView 1.6.6 [66] and saved for graphical representation with Adobe Illustrator.

S. macrospora encodes two kinases similar to mammalian STRIPAKassociated kinases STK24, STK25, MST4 and MINK1
In mammals, three members of the GCK III subfamily of the Ste20 kinases (STK24, STK25, and MST4); and one member of the GCK IV family (Misshappen-like kinase 1, MINK1) are associated with the STRIPAK complex [15,16]. To date, no STRIPAK-associated kinases have been identified in filamentous fungi (Fig 1).
Therefore, we performed a BLASTP search using mammalian kinases STK24, STK25, MST4 and MINK1 as query against the S. macrospora proteome (http://blast.be-md.ncbi.nlm.nih.gov/ [67]) to identify putative S. macrospora homologs of the mammalian kinases. This search revealed the putative germinal center kinases SMAC_01456 (F7VQV9) and SMAC_04490 (F7VYS5) to be orthologous to the mammalian GCKs ( Table 2). The predicted gene SMAC_04490 comprises 2758 bp and contains three putative introns: 96 bp at position 53-148, 117 bp at position 217-333 and 82 bp at position 502-583 ( Fig A in S1 File). The calculated molecular weight (MW) of the encoded 820-aa protein is 91.4 kDa with an isoelectric point (pI) of 9.47. It shows a high sequence identity to the serine/threonine-protein kinase 3 (stk-3) of Neurospora crassa (locus tag NCU04096, gene symbol prk-9 alias sid-1) (Fig 2) and was therefore named SmKIN3. Two variants have been reported for the N. crassa NCU04096 protein: a long version resulting from a transcript after splicing of three introns, and a short version derived from a transcript retaining intron 3 and following initiation of translation at a downstream ATG codon (http://www.broadinstitute.org/annotation/genome/neurospora/). The position of all three introns is conserved in Smkin3 and stk-3 (prk-9, sid-1), but we identified splicing of all three introns in Smkin3 using RT-PCR and cDNA sequencing, and splicing of intron 3 did not appear to be optional (Fig A in S1 File). In addition, the position of the downstream start codon is conserved in Smkin3. Initiation of translation from this ATG codon would lead to an N-terminally truncated SmKIN3 version of 663 aa with an MW of 74 kDa (Fig A in S1 File).
NCU00772 (Fig A in S1 File). Thus, in addition to the full-length protein two further variants are encoded by Smkin24. Skipping the splicing of intron 1 and initiation of translation at the downstream ATG codon would result in a 789 aa protein with a MW of 88 kDa, whereas skipping intron 4 splicing would lead to a C-terminal truncated version of 729 aa with a MW of 82 kDa (Fig B in S1 File).
However, the C-terminally located regulatory domains are structurally different to GCKs from other filamentous ascomycetes, as well as S. cerevisiae and S. pombe (less than 20%). Phylogenetic analysis revealed that SmKIN3 and SmKIN24 were strictly separated and cluster together with putative homologs from other filamentous ascomycetes (Fig 3).

SmKIN3 and SmKin24 interact physically with PRO11
Striatin is the scaffold of mammalian STRIPAK complex kinases MST4, STK24, STK25 and MINK1 [15,16]. PRO11 was previously shown to be the S. macrospora homolog of mammalian striatin [8]. The sequence similarity between SmKIN3 and SmKIN24 and the mammalian kinases of the GCK III and GCK IV family (Fig 1) led us to inspect whether both S. macrospora kinases can interact with PRO11 in a Y2H system. Full-length cDNAs (version I with all introns spliced, Fig A in S1 File) of Smkin3 and Smkin24 were cloned into the Y2H vector pGBKT7 that contains the GAL4 DNA-binding domain. The pGADT7-derived plasmid pAD11FL, which encodes a full-length PRO11 fused to the GAL4 activation domain, served as the prey vector. Plasmids were transformed into yeast strains Y187 (pGBKT7 constructs) or AH109 (pGADT7 constructs). Strains carrying pBD-SmKIN3 and pBD-SmKIN24 plasmids were checked for transactivation activity by mating with yeast strain AH109 containing pGADT7. A strain carrying pGBKT7 and pGADT7 was used as negative control. Gal4 fusion proteins of SmKIN3 and SmKIN24 were checked for adequate expression and two-hybrid competency using the RanBPM system [62]. Previously, we showed that the MOB domain protein SmMOB3, the mammalian phocein homologue, is a strong interaction partner of PRO11 [31,74]. We therefore tested the interaction of both kinases with SmMOB3. Using the Y2H system we confirmed that both SmKIN3 and SmKIN24 could physically interact with PRO11 but not with SmMOB3 (Fig 4 and Fig D in S1 File).  Shown are serial dilutions of diploid yeast strains obtained after mating and spread on SD medium lacking tryptophan (trp) and leucine (leu) or trp, leu and adenine (ade) to verify the interaction of both proteins. The left picture displays a control, which ensures presence of both plasmids in the diploid strain; trp and leu prototrophy are regained by genes present on the plasmid. The right picture displays the interaction assay; ade prototrophy is obtained by positive interaction between the proteins fused to GAL4 binding domain (BD) and GAL4 activation domain (AD). Full length versions of SmKIN3 and SmKIN24 without any introns were tested. GAL4-binding domain was fused to SmKIN3 or SmKIN24 GAL4 activation domain to PRO11. Reverse application of activation and binding domain was not possible due to transactivation of BD-PRO11. Yeast transformants expressing genes coding for BD-SmKIN3 or BD-SmKIN24 and AD-RanBPM served as positive control (46). As negative control a diploid strain carrying empty vectors pGADT7 (AD) and pGBKT7 (BD) was used. No interaction between SmMOB3 and SmKIN3 or SmKIN24 has been observed. To test whether PRO11 and GCKs SmKIN3 and SmKIN24 could interact in vivo, we performed co-IP studies in S. macrospora. We expressed N-terminally FLAG-tagged SmKIN3 and HA-tagged PRO11 [31] in S. macrospora. We were not able to express FLAG-tagged SmKIN24 in S. macrospora wt in amounts suitable for co-IP, under control of either the native or the strong, constitutive ccg1 promoter of N. crassa [61]. Expression of Smkin24 in the ΔSmkin24 deletion background was possible in low amounts suitable for localization (see below) but expression in wt for co-IP was necessary since only two resistance markers exist and no recombinase-mediated excision system is yet established for S. macrospora. All obtained viable transformants expressed Smkin24 in low amounts. We therefore concluded that SmKIN24 is highly unstable. However, tagged SmKIN3 and PRO11 could be expressed separately and coexpressed in wt S. macrospora. Pull-down experiments without adding a crosslinker did not result in precipitation of PRO11 (Fig E in S1 File). Since protein kinases often interact weakly and transiently with their substrates, we performed sulfhydryl crosslinking with BMH which confirmed the physical interaction of full length SmKIN3 and PRO11 (Fig 5).

Deletion of Smkin3 or Smkin24 impairs vegetative growth but only ΔSmkin24 is sterile
In order to investigate the functional role of Smkin3 and Smkin24, we generated single-and double-deletion strains ΔSmkin3, ΔSmkin24 and ΔSmkin3/ΔSmkin24 (Fig F and Fig G in S1 File). The respective genes were replaced with a hygromycin resistance cassette by homologous recombination in the S. macrospora Δku70 [54] strain (Fig F and Fig G in S1 File). The Δku70 background was removed afterwards by crossing the homokaryotic deletion strains with S. macrospora spore-color mutant r2 [3] and the obtained single-spore isolates were verified by PCR and Southern blotting (Fig F and Fig G in S1 File). Microscopic analysis of the knockout strains revealed that only ΔSmkin3 was fertile and able to generate fruiting bodies with ascospores within 7 days (Fig 6). The deletion strain ΔSmkin24 was halted during late protoperithecia maturation and thus failed to develop mature fruiting bodies (Fig 6). Additionally, ΔSmkin3 produced more aerial hyphae on solid SWG fructification medium compared to wt (Fig 6). The double-deletion strain was generated by crossing the single deletion strains, and ΔSmkin3/ΔSmkin24 exhibited a phenotype that was a combination of the phenotypes of both single-deletion strains (sterility and increased aerial hyphae on solid SWG medium). Each of the described phenotypes could be complemented by inserting a copy of the deleted gene ectopically (Fig H in S1 File). Sterility of STRIPAK mutants pro11, pro22, Δpro45 and ΔSmmob3 is accompanied by defects in hyphal fusion [31,32,74]. However, in contrast with these previously characterized mutants, fusion of vegetative hyphae was not affected in ΔSmkin3 and ΔSmkin24 (Fig I in S1 File).

SmKIN3 and SmKIN24 localize to septa and influence septum formation
In order to determine the localization of SmKIN3 and SmKIN24 in vivo, fluorescence microscopy was performed. Since C-terminally tagged versions of SmKIN3 and SmKIN24 did not complement the mutant phenotype, genes encoding N-terminally eGFP-tagged full-length SmKIN3 or SmKIN24 were expressed in the respective S. macrospora deletion strains. These constructs complemented the mutants and revealed that both kinases were localized at the septa (Fig 7). However, SmKIN3 was localized mainly at the outer part of the septum, whereas SmKIN24 was localized at the septal pore (Fig 7), and this was verified by co-staining with Calcofluor white (Fig 7). These findings lead us to quantify septum formation in the ΔSmkin3, . FLAG-SmKIN3 was pull-downed with HA-PRO11 as bait and identified by Western blot using anti-FLAG antibody (IP-HA). For this experiment, a strain expressing FLAG-SmKIN3 and the HA tag without any fused protein served as negative control and was treated equivalent to the co-IP sample. To exclude unspecific antibody binding, protein extract of the wt was used as second negative control for Western blot analysis. HA-PRO11 was isolated using FLAG-SmKIN3 as bait and identified by Western blot (IP-FLAG) using an anti-HA antibody. A strain expressing genes coding for the FLAG-tag without any fused protein and HA-PRO11 served as negative control and was treated as the co-IP sample. To exclude unspecific antibody binding, protein extract of the wt was used as second negative control for Western blot analysis. "+" mark strains expressing the respective construct.  ΔSmkin24 and ΔSmkin3/ΔSmkin24 strains and their respective complementation strains by staining with Calcofluor white after 18 or 32 h of growth ( Fig 8A). Furthermore, we quantified the distances between adjacent septa in wt and deletion mutants after 24 h of growth (Fig 8B). Septa were distributed in a uniform manner in wt with hyphal compartment length between 31-70 μm (Fig 8B). In contrast, ΔSmkin3 exhibited larger distances between adjacent septa (>71 μm), although this effect begun to revert after 24 h of growth ( Fig 8A). Strikingly, ΔSmkin24 developed numerous closely-packed septal bundles of abnormal shape, with much smaller hyphal compartments of 0-30 μm (Fig 8B). ΔSmkin3 produced about half of the total number of septa within a distance of 18 mm than wt, while ΔSmkin24 produced 20% more than wt after 24 h of growth. With respect to septation, the double-deletion mutant ΔSmkin3/ ΔSmkin24 displayed a similar phenotype to the ΔSmkin24 single-deletion strain. Ectopic integration of the respective wt gene into deletion mutants restored septal formation to wt level ( Fig  8A and 8B). The septa of ΔSmkin3 and ΔSmkin24 were checked for functionality under the microscope directly after cutting of the hyphae. Both strains did not differ from wt; shortly after cutting, hyphal compartments of ΔSmkin3 and ΔSmkin24 were plugged by Woronin bodies.

ΔSmkin3 protoplasts recover significantly faster than wt protoplasts
We observed an increased growth rate of protoplasts from mutant ΔSmkin3 compared to wt protoplasts, therefore we isolated protoplasts from wt, and strains ΔSmkin3, ΔSmkin24 and ΔSmkin3/ΔSmkin24 and their respective complementation strains using a previously described  protocol [50,51] and adjusted the protoplast concentration to 4 x 10 4 protoplasts/ml. The protoplasts were spread on solid complete medium with 10.8% saccharose (CMS) [50] and microscopically and macroscopically analyzed after 24, 48 and 72 h of growth (Fig 9). After 24 h, wt protoplasts developed a slightly branched mycelium, which expanded further by 48 h (Fig 9). Mycelia were only faintly visible without magnification on agar plates after 72 h. Regeneration of protoplasts from strain ΔSmkin24 resembled those of wt. In contrast, protoplasts from mutants ΔSmkin3 and ΔSmkin3/ΔSmkin24 recovered markedly faster within the first days than wt and mycelia generated by mutant ΔSmkin3 protoplasts were much denser and clearly visible with the naked eye after 72 h. Increased aerial hyphae were also present in mutant ΔSmkin3 and this effect was complemented by ectopically integrated wt copy of Smkin3 in the ΔSmkin3 mutant (Fig 9). The accelerated growth rate of ΔSmkin3 vegetative mycelium disappeared after five days. The increase distance between adjacent septa of mutant ΔSmkin3 could also cause larger protoplasts which are more likely to recover. Therefore, we measured the diameter of protoplasts from wt, mutants ΔSmkin3 and ΔSmkin24, but observed no differences between all strains. Moreover, addition of the cell wall disturbing drug Calcofluor white did not affect the increased growth rate of hyphae derived from ΔSmkin3 protoplasts.

Discussion
In mammals, GCKs are implicated in fundamental developmental processes, such as proliferation, cell polarity and cell migration [40,[75][76][77]. Previously, it was shown that the GCKs MST4, STK24, STK25 and MINK1 bind to striatin and therefore are part of the mammalian STRIPAK complex [15,16]. The STRIPAK core components striatin and phosphatase PP2A are conserved from yeast to man [37] (Fig 1A). The S. macrospora STRIPAK complex contains the striatin homologs PRO11 and PP2A catalytic and scaffold subunits as well as the STRIP1/2 homolog PRO22, the SLMAP homolog PRO45 and the phocein homolog SmMOB3 [31,32]. However, until now, evidence for the presence of STRIPAK-related kinases in ascomycetes was lacking. Here, we analyzed putative GCKs in S. macrospora and addressed their relevance as putative STRIPAK-associated kinases.

Are SmKIN3 and SmKIN24 STRIPAK-associated kinases?
A BLASTP search identified SmKIN3 and SmKIN24 as putative homologs of mammalian STRIPAK kinases. The N-terminal part of the mammalian STRIPAK kinases, containing the kinase domain, is conserved among ascomycetes (Fig 2). Moreover, the phylogenetic analysis of MST4, STK24, STK25 and MINK1 and their orthologs in ascomycetes revealed two groups of putative STRIPAK-associated GCKs in filamentous fungi (Fig 3).
N. crassa STK-3 (PRK-9, SID-1) and STK-24 (STK-6, MST-1) both localize to the septa and spindle pole bodies (SPB) [35,78]. In S. pombe, SIN components bind to the coiled coil scaffold protein SID4 [81]. However, a clear homolog of the SID4 has not been identified in ascomycetes [78]. Based on interaction studies, STRIPAK PRO11 may serve as a scaffold for SmKIN3 and SmKIN24 and we therefore predict interplay between STRIPAK and SIN/MEN in S. macrospora. This hypothesis is supported by findings in fission yeast, where the STRIPAK-like complex is called SIP complex (Fig 1A) [28]. The fission yeast SLMAP-homolog CSC1 was shown to negatively regulate SIN and promote SIN asymmetry [28] which is crucial for cytokinesis. In D. melanogaster, STRIPAK components negatively regulate Hippo signaling [20], which is homologous to the fungal SIN/MEN and RAM/MOR networks [23,82] (Fig 1B). The Hippo pathway is a central element of tissue size control in D. melanogaster and higher organisms. It consists of an upstream kinase Hpo (MST1/2 in mammals), downstream NDR kinases LATS1/2 and NDR1/2 associated MOB family proteins and various scaffolding subunits [23,83,84] (Fig 1B). Based on interaction of Hpo/MST1/2 with homologs of MOB3 and striatin, the D. melanogaster dSTRIPAK complex was identified as the negative regulator of Hippo signaling. Loss of dSTRIPAK components resulted in elevated phosphorylation of Hpo [20]. Similarly, in mammals it was demonstrated that the Hippo kinases MST1 and MST2 bind the SLMAP-STRI-PAK complex [21]. A BLASTP analysis with Hpo, MST1 and MST2 protein sequences revealed SmKIN3 and SmKIN24 as orthologs of these kinases in S. macrospora. However, SmKIN3 and SmKIN24 did not interact with SmMOB3 in Y2H experiments (Fig 2). In N. crassa, a direct interaction of STK-24 (SID-1) and the NDR kinase DBF-2 was identified by MS and yeast twohybrid studies; however an interaction with the kinase activator MOB-1 was not tested. It may be therefore meaningful considering interactions of other MOB-family proteins and STRIPAK associated GCKs to define the regulatory network of the STRIPAK complex and the Hippo (SIN/MEN-MOR/RAM) pathways. Dettman et al. [85] identified an interaction between STK-24 (STK-6, MST-1), the N. crassa homolog of SmKIN24, and components of the MAK-2 mitogen activated protein kinases module, (MAPKKK) NRC-1 and (MAPKK) MEK-2, involved in fungal communication between germinating spores [86,87]. All three kinases of the MAK-2 MAPK module were shown to be weakly associated with STRIPAK components PP2A-A, PP2A-C, striatin and MOB-3 in N. crassa [85]. Thus, STRIPAK components and the SmKIN24 homolog function together in at least one cascade in N. crassa. In mammals, CCM3 recruits the GCKs MST4, STK24 and STK25 but not MINK1 to striatin [15,17,45]. However, a homolog of CCM3 has so far not been identified in fungi. Thus, we assume that STRIPAK-associated kinases in S. macrospora are not recruited by other proteins to PRO11.
SmKIN3 and SmKIN24 affect growth rate, sexual development and septum formation Deletion of Smkin3 does not impair fertility, but leads to an increased formation of aerial hyphae and reduced growth rate. In contrast, the deletion strain ΔSmkin24 was sterile, its lifecycle was halted at the later stage of protoperithecia formation and vegetative growth was also impaired (Fig 6). A global phenotypic analysis of serine/threonine-protein kinase deletion mutants in N. crassa revealed that the N. crassa Smkin3-homolog stk-3 (prk-9, sid-1) affects asexual and sexual development as well as vegetative growth and aerial hyphae formation [88]. S. macrospora undergoes only sexual development, so we cannot comment on asexual reproduction in this study. The reduced growth rate of deletion strain ΔSmkin3 confirmed previous results of Park et al. [88], however, in contrast to these findings, deletion of Smkin3 did not impair fruitingbody development or fertility. Heilig et al. [78] reported that N. crassa wt x Δstk-3 (prk-9, sid-1) crosses exhibited no abnormalities during sexual development. Deletion of stk-24 (stk-6, mst-1) in N. crassa resulted in reduced aerial hyphae and slightly reduced macroconidia production [89]. Similar to the S. macrospora ΔSmkin24 mutant which is halted at later stages of protoperithecia formation, N. crassa Δstk-24 (stk-6, mst-1) x Δstk-24 (stk-6, mst-1) crosses resulted in empty perithecia that do not contain ascospores. Unlike S. macrospora mutant ΔSmkin24, the mycelial extension rates were not impaired in the N. crassa Δstk-24 (stk-6, mst-1) mutant [35].
Both N. crassa Δstk-3 (prk-9, sid-1) and Δstk-24 (stk-6, mst-1) mutants showed hyphal tip swelling and bursting of hyphal tips followed by cytoplasmic leakage [78,89]. This phenotype was not observed in the S. macrospora ΔSmkin3 and ΔSmkin24 mutants. Deletion of Smkin3 and Smkin24 cause defects in septation, and the SmKIN3 and SmKIN24 proteins are localized to the septa (Fig 7). In N. crassa, the respective homologs are also localized to the septum and deletion of the genes causes defects in septa formation [35,78,89].
Mutant ΔSmkin3 exhibited fewer septa compared with the wt strain, and hyphal compartments were therefore elongated (Fig 8). This phenotype reverted, as growth continued beyond 24 h, presumably due to the accumulation of suppressor mutations. This is similar to the N. crassa Δstk-3 (prk-9, sid-1) mutant that produces initially aseptate germlings that form septa during the later stages of colony development [78]. In addition, germlings derived from ΔSmkin3 protoplasts grew faster than germlings of wt protoplasts (Fig 9). The septation phenotype of mutants ΔSmkin24 and ΔSmkin3 were distinct, mutant ΔSmkin24 generated more septa that aggregated into bundle-like structures of abnormal shape (Fig 8). Septal actomyosin tangle (SAT) assembly, cortical actomyosin ring (CAR) assembly and CAR constriction are three consecutive stages of septum formation in N. crassa [90]. The septa observed in mutant ΔSmkin24 resembled unfinished septa from the early stages of septum formation, that were observed during MYO-2 GFP localization studies of SAT and CAR assembly in N. crassa [90]. SmKIN24 is therefore required for proper CAR assembly, as was recently shown by Heilig et al. [35] in N. crassa. Despite their contrary phenotypes, SmKIN3 and SmKIN24 appear to function independently in S. macrospora, as shown by the double-deletion mutant ΔSmkin3/ ΔSmkin24 that exhibited a phenotype that combined the phenotypes of both single-deletion strains, although septum formation resembles that of the ΔSmkin24 mutant (Fig 8).
The N. crassa homologs of SmKIN3 and SmKIN24 are implicated in SIN and MOR pathways as discussed above. The results of the present study indicate an additional role of these GCKs in S. macrospora, potentially via interaction with the striatin homolog PRO11. The interaction between PRO11 and SmKIN3 or SmKIN24 reported here for the first time, should be confirmed by other in vivo methods, such as bimolecular fluorescence complementation coupled with high resolution microscopy, which could also provide more detailed localization information of a possible interaction. In addition, interactome studies could identify transient interaction partners and may give insights into crosstalk between signaling pathways and modulation by protein complexes.
Supporting Information S1 File. Table A, Plasmids used for this study. Table B, Primers used for this study.  Fig D, Controls of yeast two-hybrid assay AD-PRO11 and AD-SmMOB3 with empty vector pGBKT7. Shown are serial dilutions of diploid yeast strains obtained after mating and spread on SD medium lacking tryptophan (trp) and leucine (leu) or trp, leu and adenine (ade) to test the interaction of both proteins. The left picture displays a control, which ensures presence of both plasmids in the diploid strain; trp and leu prototrophy are regained by genes present on the plasmid. The right picture displays the interaction assay; ade prototrophy is not obtained by negative interaction between the GAL4 binding domain (BD) and proteins PRO11 and SmMOB3 fused GAL4 activation domain (AD).  [46]. Positions of the respective probes are indicated in (A). ΔSmkin24 was verified using a probe binding at the 3' region of Smkin24. The successful integration is represented by a band shift. Fig H, Macroscopic and microscopic analysis of the sexual development of wt, complemented ΔSmkin3 (ΔSmkin3+), complemented ΔSmkin24 (ΔSmkin24+) and partially complemented ΔSmkin3/ΔSmkin24. The wt strain produces ascogonia after 3 days which develop to unpigmented protoperithecia at day 4 and pigmented protoperithecia at day 5. After 7 days mature perithecia with asci and ascospores are formed. Similar to the wt ΔSmkin3+ and ΔSmkin24+ and ΔSmkin3/ΔSmkin24 + Smkin24 completed the lifecycle within 7 days and produced mature ascospores. Development of ΔSmkin3/ΔSmkin24 + Smkin3 is arrested at stage of late protoperithecia formation, similar to ΔSmkin24. Scale bars as indicated. Fig I, Microscopic investigation of hyphal fusion in wt, ΔSmkin3, ΔSmkin24 and ΔSmkin3/ΔSmkin24. ΔSmkin3, ΔSmkin24 and ΔSmkin3/ΔSmkin24 are capable of hyphal fusion. Hyphal fusion events are highlighted with circles. Pictures of hyphal fusion events were taken at subperiphal regions 10 mm behind the growth front. Hyphal fusion was investigated 2-3 days past inoculation. (PDF)