The GCKIII Kinase Sps1 and the 14-3-3 Isoforms, Bmh1 and Bmh2, Cooperate to Ensure Proper Sporulation in Saccharomyces cerevisiae

Sporulation in the budding yeast Saccharomyces cerevisiae is a developmental program initiated in response to nutritional deprivation. Sps1, a serine/threonine kinase, is required for sporulation, but relatively little is known about the molecular mechanisms through which it regulates this process. Here we show that SPS1 encodes a bona-fide member of the GCKIII subfamily of STE20 kinases, both through phylogenetic analysis of the kinase domain and examination of its C-terminal regulatory domain. Within the regulatory domain, we find Sps1 contains an invariant ExxxPG region conserved from plant to human GCKIIIs that we call the EPG motif; we show this EPG motif is important for SPS1 function. We also find that Sps1 is phosphorylated near its N-terminus on Threonine 12, and that this phosphorylation is required for the efficient production of spores. In Sps1, Threonine 12 lies within a 14-3-3 consensus binding sequence, and we show that the S. cerevisiae 14-3-3 proteins Bmh1 and Bmh2 bind Sps1 in a Threonine 12-dependent fashion. This interaction is significant, as BMH1 and BMH2 are required during sporulation and genetically interact with SPS1 in sporulating cells. Finally, we observe that Sps1, Bmh1 and Bmh2 are present in both the nucleus and cytoplasm during sporulation. We identify a nuclear localization sequence in Sps1 at amino acids 411–415, and show that this sequence is necessary and sufficient for nuclear localization. Taken together, these data identify regions within Sps1 critical for its function and indicate that SPS1 and 14-3-3s act together to promote proper sporulation in S. cerevisiae.


Introduction
Yeast deprived of a fermentable carbon source and nitrogen undergo sporulation [1].Sporulation begins with meiosis, which results in the production of four haploid nuclei from a single diploid cell.These four nuclei are encapsulated by the prospore membrane, which acts as the template for spore wall deposition.The spore wall differs from the vegetative cell wall, and contains the spore-specific chitosan and dityrosine layers that protect the spores during times of harsh environmental stress.Sporulation is a highly regulated process, and SPS1, which encodes a STE20 family serine/threonine kinase, is essential for sporulation [2].STE20 family kinases are highly conserved from yeast to mammals and are divided into two subgroups, the p21-activated kinases (PAKs) and the germinal center kinases (GCKs) [3,4].These two subgroups are distinguished both by the phylogenetic relationships among their kinase domains and by their domain architectures: In PAKs, the kinase domain is C-terminal to the regulatory domain, and this is reversed in GCKs [5].Within the GCKs, the GCKIII subfamily of kinases includes the mammalian kinases MST3, MST4, and YSK1/SOK1/STK25 [3], which have been implicated in processes such as apoptosis [6] and axon outgrowth [7], and may be involved in diseases such as Alzheimer's [8], type 2 diabetes [9], Parkinson's disease [10], and cerebral cavernous malformations [4].
In S. cerevisiae, SPS1 is required for proper sporulation.In particular, previous work has shown that SPS1 is required for the proper localization of the Gsc2, Chs3, and Gas1 enzymes involved in the construction of the spore wall [2,11,12].In addition, Sps1 may play a role in histone modification [13], although whether this role is direct is currently unclear.SPS1 has also been shown to regulate yeast replicative lifespan [14].
At the molecular level, 14-3-3 proteins are acidic, readily form dimers and bind other proteins using a conserved binding groove [28].Binding by 14-3-3 proteins has been shown to affect protein function through multiple mechanisms which include acting as a scaffold to facilitate interaction between proteins, modulating protein degradation rate, and altering protein subcellular localization [29].14-3-3 binding to substrates in a phosphorylation dependent manner was first shown between 14-3-3f and a serinephosphorylated Raf-1 peptide [30].Subsequently three different consensus sequences for 14-3-3 binding have been identified: RSX(pS/pT)XP, RXXX(pS/pT)XP [31] and (pS/pTX)(1-2)-COOH [32] (where pS/pT indicates a phosphoserine or phosphothreonine respectively and X represents any amino acid).
In this study, we use phylogenetic analysis to determine the relationship of Sps1 to other Ste20 kinases, and demonstrate that Sps1 is a bona-fide member of the GCKIII family of STE20 kinases.Our comparative analyses also identify a C-terminal region in GCKIII kinases that is conserved from yeast to mammal to plant, and we show that this region is important for Sps1

Plasmids used in this study
All plasmids used in this study can be found in Table S1 and all primers in Table S2.Construction details are described below.All plasmid inserts amplified using PCR were verified by sequencing.
pCS47 (pBSIIKS+: 2137 SPS1 +297 ) was constructed by amplifying genomic DNA using primers OLH778 and OLH1195.The resulting product, as well as pBSIIKS+, were cut with the restriction enzymes ClaI and SpeI followed by ligation.

Strains used in this study
All strains in this study were derived from the SK1 background [50] (list of strains, Table S3).C-terminal tagging, gene disruptions, and gene deletions were accomplished by PCR mediated recombination [51,52].All of the above genome changes were confirmed by diagnostic PCR.

Fluorescence microscopy
Microscopy was done using a 1006 (NA 1.45) objective on a Zeiss Axioskop Mot2.Images were taken using an Orca-ER cooled charge-coupled device camera (Hamamatsu) using Openlab 4.04 (Perkin Elmer) software.All imaging was done using live cells.

Yeast growth conditions
Yeast cells were induced to sporulate as described previously [50].In brief, yeast cells were grown to saturation in YPD (2% Peptone, 1% Yeast Extract, 2% Dextrose), and transferred to presporulation media YPA (2% Peptone, 1% Yeast Extract, 1% Potassium Acetate).Cells were grown in pre-sporulation media overnight, and then shifted to sporulation media (2% Potassium Acetate).Cells grown in log phase were grown in either YPD or the appropriate selective medium to approximately OD 600 0.8.

Scoring of sporulation phenotypes
Spore efficiency was measured using liquid sporulation cultures from a minimum of three biological replicates.Meiotic progression was determined by examining Htb2-mCherry [48].Cultures were allowed to sporulate in liquid media for 24 hours.Yeast cells were scored for sporulation if at least one refractile spore was formed; a minimum of 200 cells was counted for each culture.
To determine the number of spores formed in each ascus, three biological replicates were tested for each strain.Cultures were sporulated in liquid sporulation media for 24 hours.Yeast cells were scored for number of refractile spores formed per ascus.Cells that did not form refractile spores were not counted.
The dityrosine fluorescence assay was performed as described [53].In short, cells were grown on YPD plates for 24 hours then transferred to SPO plates with a nitrocellulose filter, incubated for another 24 hours and then exposed to UV light.The nitrocellulose membrane was then imaged using a digital camera.
Spore viability was assayed by the dissection of 25 tetrads per strain [54].Dissected spores were allowed to grow on YPD plates for 48 hours.
Spore wall permeability assays were carried out as previously described [55].In brief, yeast strains were transformed with pRS424-ssGFP, induced to sporulate, and visualized for GFP in the ascal cytoplasm or trapped within the extracellular space between the plasma membrane and the spore wall.For each strain 100 cells were counted from three biological replicates for a total of 300 total cells counted per strain.Only cells with refractile spores were counted.
Blots visualized on the Kodak Image Station 4000R with Kodak Molecular Imaging Software v4.0.4 were stained with the appropriate anti-rabbit or anti-mouse HRP conjugated secondary antibodies at 1:10000 (Jackson ImmunoResearch).Secondary antibodies detected using Supersignal West Dura Extended Duration Substrate (Pierce).
Blots visualized on the Odyssey CLx Infrared Imaging System (LI-COR Biosciences) were blocked using Odyssey blocking buffer (LI-COR) and Goat Anti-Mouse IR Dye 800 CW (LI-COR) at 1:10000 as a secondary antibody.Protein bands were quantified using Image Studio v3.1 (LI-COR).
Sps1-zz immune complexes were washed 4 times in IP buffer and then re-suspended in SDS-PAGE sample buffer, boiled for 5 minutes and separated by SDS-PAGE.SBP-Sps1 immune complexes were washed 4 times in IP buffer and then SBP-Sps1 was eluted using biotin solution (2 mM biotin (Sigma), 0.28% ammonium hydroxide (Sigma)).Eluate was then TCA prepped as described above.The TCA precipitated protein was then resuspended in SDS-PAGE sample buffer, boiled for 5 minutes, and separated by SDS-PAGE.
Reactions were incubated for 30 minutes at 30uC.Beads were then resuspended in SDS-PAGE sample buffer, boiled for 5 minutes and separated by SDS-PAGE.
Samples for mass spectrometry phosphorylation analysis were immunoprecipitated and then resuspended in SDS-PAGE sample buffer, boiled for 5 minutes and separated by SDS-PAGE.Gel bands were excised and sent to the Taplin Mass Spectrometry Facility (Harvard Medical School).In brief, samples underwent a modified in-gel trypsin digestion [57], followed by peptide extraction procedures, separation on a reverse-phase HPLC capillary column [58], electrospray ionization and peptides then entered into an LTQ-Orbitrap mass spectrometer (Thermo Fisher, San Jose, CA).Peptide sequences were determined by matching protein or translated nucleotide databases with the acquired fragmentation pattern by the software program, Sequest [59].To identify phosphorylated residues, the modification of 79.9663 mass units to serine, threonine, and tyrosine was included in the database searches to determine phosphopeptides and each phosphopeptide that was identified by the Sequest program was also inspected manually.Bioinformatics ScanSite, http://scansite.mit.edu/[60], was used to predict the 14-3-3 binding site in Sps1.ScanSite was run using the high stringency setting.

Sps1 is a GCKIII family member
Sps1 has been considered an outlier among Ste20 kinases, as initial analysis of the evolutionary relationships among Ste20 family kinases suggested it belonged to neither GCK nor PAK subfamilies [5].However, we noticed that when compared to human Ste20 kinases, the kinase domain of Sps1 shares ,50% amino acid identity with the GCKs (e.g., 52% identity with the human GCKIII Ysk1), but only ,40% with the PAKs (e.g., 40% with the human Pak1).This raised the possibility that rather than being an outlier among Ste20 kinases, Sps1 might belong to the GCK subfamily, and prompted us to revisit the evolutionary relationship of the Sps1 kinase domain to other Ste20 family members.
Phylogenetic analysis of Sps1 was performed to examine relationships among the kinase domain sequences of all five S. cerevisiae Ste20 family members (Cla4, Kic1, Skm1, Sps1, and Ste20), as well as all human, Drosophila melanogaster and Arabidopsis thaliana Ste20 family members.Two MAP Kinase Kinases (MEKKs), S. cerevisiae Ste11 and human MEKK1, were used to root the tree, as MEKK and Ste20 kinases form related but distinct gene families (Figure 1).Maximum likelihood-based phylogenetic analysis found, with strong branch support, that Sps1 resides in a monophyletic clade with human Mst3, Mst4, and Ysk1, as well as Drosophila GCK-III.As Mst3, Mst4, Ysk1, and GCK-III are canonical GCKIII family members [3][4][5], these data indicate that Sps1 is not an outlier among Ste20 kinases.Our analysis shows that Sps1 is a GCK, and specifically, a member of the GCKIII subfamily.Phylogenetic analysis performed using a different reconstruction approach (the less computationally intense Neighbor Joining method) also indicated Sps1 belongs to the GCKIII family.Consistent with this view, the Sps1 domain architecture also resembles GCK architecture, with the kinase domain located at the amino-terminus.Interestingly, two Arabidopsis MAP4Ks also fall within the GCKIII clade, suggesting that the GCKIII kinase lineage diverged from other Ste20 kinases prior to the separation of the plant and opisthokont (yeast/animal) lineages ,1 billion years ago.The other yeast Ste20 family member with a GCK-like domain architecture, Kic1, is also a member of the GCKIII clade.
As a member of the GCKIII family, Sps1 could share amino acid similarities outside the kinase domain.We performed sequence alignments with Sps1 and other GCKIIIs and found Sps1 contains a conserved region at its C-terminus (Figure 2).This region extends from amino acid 453 to 482 in Sps1, and it includes three amino acids conserved between Sps1 and the animal and plant GCKIIIs (Figure 2).In Sps1, these residues are glutamic acid (E) 464, proline (P) 468, and glycine (G) 469.We call this invariant ExxxPG region the EPG motif.Taken together, the phylogenetic evidence, domain architecture, and C-terminal sequence similarity all support the identification of Sps1 as a member of the GCKIII subfamily of Ste20 kinases.

The C-terminal EPG motif containing region is required for Sps1 function
To test whether the C-terminal EPG motif containing region is required for Sps1 function, we created the sps1DEPG-zz allele by removing the last 38 amino acids (and thus deleting the ExxxPG region).The deletion starts from the conserved valine (V) 453 to the C-terminus; this sequence is replaced with the zz (two tandem z domains from Protein A) epitope [52].We compared a strain carrying the sps1DEPG-zz allele to strains carrying the alleles: SPS1-zz, sps1::LEU2 as well as wild type for their ability to form refractile spores (summarized in Table 1).We found that wild type and SPS1-zz cells sporulate at 80.7% and 81.7% respectively, while sps1DEPG-zz reduces sporulation similar to the sps1 null allele (6.8% and 4.0%, respectively) (Student's t test comparison of LH960 and LH961 gives a P,0.01).This reduction in sporulation is not due to a reduction in protein level, as the sps1DEPG-zz mutation does not grossly disrupt steady state levels of Sps1 protein (Figure S1A).

Sps1 is a phosphoprotein
We examined Sps1 during sporulation and saw that the level of Sps1 is induced during sporulation and peaks around 6 to 8 hours (Figure 3A), when cells are completing meiosis and starting spore morphogenesis.We see that Sps1 runs as a doublet, suggesting that Sps1 is post-translationally modified.
To test whether this post-translational modification was due to phosphorylation, we carried out a phosphatase assay by examining Sps1 purified from cells 8 hours into sporulation (Figure 3B).When l phosphatase was added to the immunoprecipitated Sps1 we see that the more slowly migrating band is no longer detectable.When l phosphatase and phosphatase inhibitors were added together, the higher mobility band can be seen more easily than in the phosphatase only treated sample (Figure 3B).These results indicate that Sps1 is phosphorylated (Figures 3A and 3B).Differences in the intensity of the more slowly migrating phosphorylated band in the phosphatase treatment experiment (Figure 3B) compared to the examination of cell lysates (Figure 3A) may be due to the use of different sample preparations (more native conditions for the phosphatase assay in Figure 3B compared to denaturing conditions for the cell lysates in Figure 3A) or the different epitope tags used in the two different experiments.
To identify potential phosphorylation sites on Sps1, we used mass spectrometric analysis on Sps1 purified from vegetatively growing log-phase cells that expressed Sps1 from a multi-copy plasmid under the TEF2 promoter and from sporulating cells.Kinase dead sps1 was also purified from log-phase cells and subjected to mass spectrometric analysis.We found that Sps1 was phosphorylated on threonine (T) 12 and serine (S) 345 in all three experiments.In addition to T12 and S345, four novel phosphorylation sites were found only when the kinase dead version of Sps1 was purified: S19, S22, S312, and S329 (Figure 3C).Thus, Sps1 is phosphorylated on multiple serine and threonine residues, and the phosphorylation of T12 and S345 are not simply due to autophosphorylation by Sps1.Whether S19, S22, S312, or S329 are biologically relevant is unclear, given that we only see these phosphorylation sites on the kinase dead version of Sps1 that is ectopically expressed in log phase cells.

sps1-T12A displays reduced sporulation efficiency
To test the importance of Sps1 T12 or S345 phosphorylation during sporulation, we replaced the SPS1 locus with either SBP-sps1-T12A or SBP-sps1-S345A.We did not see any sporulation defect with the SBP-sps1-S345A mutant and did not study the  effects of this site further.In contrast, the SBP-sps1-T12A did affect spore formation.We found that the sps1 null forms spores 4.0% of the time, while wild type and the SBP-SPS1 strain sporulate at 80.7% and 85.8% of the time respectively.SBP-sps1-T12A forms spores 65.7% of the time, a statistically significant decrease in sporulation efficiency compared to wild type (P,0.01;t test), although not as pronounced as in an sps1 null (Table 2).
To examine whether the T12A mutation disrupted Sps1 protein levels or altered its temporal expression profile, we compared wild type SBP-Sps1 and SBP-Sps1-T12A expression during sporulation and found comparable protein levels and similar temporal expression patterns throughout sporulation (Figure S1B).

sps1-T12A is a reduction-of-function mutation defective in spore packaging
Because the SBP-sps1-T12A allele did not have a phenotype as severe as the null allele, we wanted to determine the manner in which is disrupts SPS1 function.First, we tested whether SPB-sps1-T12A acts as a dominant negative mutant by creating a heterozygous strain.If sps1-T12A were acting as a dominant negative mutation, we would expect a reduction in spore efficiency in heterozygous cells compared to wild type cells.However, we found that there was no reduction in spore formation between wild type (80.7%) and SBP-sps1-T12A/+ (82.8%;P = 0.4, Student's t test).
To determine whether the spores that are produced in sps1-T12A allele show defects, we examined the number of spores formed per ascus when at least one refractile spore was formed in an ascus (Figure S2A), the ability of the outer spore wall layer (the dityrosine layer) to fluoresce [53] (Figure S2B), as well as spore wall permeability using a signal sequence fused to GFP (pRS424-ssGFP; [55]) (Figure S2C).We saw no differences in the spores formed by cells homozygous for the sps1-T12A allele compared to wild type cells.We also examined the germination efficiency in SBP-sps1-T12A cells by dissecting 25 tetrads each of SBP-sps1-T12A and SBP-SPS1 (only cells that made four refractile tetrads in an ascus were dissected) and did not see any differences in the  ability of haploid spores to germinate.Of 100 spores dissected, 98 germinated in the SPB-SPS1 containing strain, while 96 germinated in the SBP-sps1-T12A containing strain.Taken together, these results suggest that while sps1-T12A reduces the efficiency of spore packaging, the spores that are packaged and form refractile structures show no obvious defects.
Sps1 has a consensus 14-3-3 binding site and can physically interact with Bmh1 and Bmh2 When we examined the region surrounding T12 within the Sps1 protein, we found that this threonine resides within a potential 14-3-3 phosphopeptide binding consensus sequence.The sequence of Sps1 starting from arginine (R) 9 to proline (P) 14 closely matches the 14-3-3 consensus sequence: R-S-X-(pS/pT)-X-P [31] (Figure 4A).
To examine Bmh1 expression in the SK1 strain, we used an anti-Bmh antibody [35].To test the specificity of the anti-Bmh antibody in the SK1 strain, we examined its ability to recognize Bmh1 and Bmh2 using wild type, bmh2D, BMH1-GFP, and BMH2-GFP containing strains, which allowed us to distinguish between the two 14-3-3 isoforms, as appending GFP at the Cterminus of the protein shifts the migration of an isoform (Figure 4B).We found that this antibody can detect both isoforms of 14-3-3 proteins in our strain background, with the more slowly migrating isoform corresponding to Bmh2 while the faster migrating isoform is Bmh1, as expected by their predicted protein sizes (31 kDa for Bmh2 and 30 kDa for Bmh1).
We next asked if Sps1 could bind to Bmh1 and Bmh2, and found that we can co-immunoprecipitate Bmh1 and Bmh2 with SBP-Sps1 (Figure 4C).However, we do not detect an interaction of Bmh1 and Bmh2 with SBP-Sps1-T12A (Figure 4C).These results indicate that Sps1 interacts with Bmh1 and Bmh2 and that T12 within Sps1 is required for the interaction.

Bmh1 and Bmh2 are expressed during sporulation and are present in the nucleus and the cytoplasm
We examined Bmh1 and Bmh2 expression during sporulation in wild type cells, and found that both are expressed throughout sporulation (Figure 5A).As Bmh1 was previously reported to be the major isoform in vegetative growth [35,65], we compared the levels of expression of Bmh1 and Bmh2 during log-phase growth and sporulation.We used the strain BMH1-GFP BMH2-GFP and examined the protein levels using an anti-GFP antibody to avoid any bias in isoform detection by the anti-Bmh antibody.We found that the ratio of Bmh1 to Bmh2 in log-phase growth was 3.460.2(mean 6 S.D.), decreasing to 1.760.2(mean 6 S.D.) in sporulating cells.This suggests that Bmh1 is more abundant than Bmh2 during log-phase growth, but that Bmh2 levels increase relative to Bmh1 in sporulating cells (Figure 5B).
We also examined Bmh1 and Bmh2 localization during sporulation.High-throughput studies have examined Bmh1 and Bmh2 in haploid log-phase cells and found both cytoplasmic and nuclear localization [66].We examined Bmh1-GFP localization in diploids during both log-phase growth and sporulating cells and saw both nuclear and cytoplasmic localization (Figure 5C, top).We see similar results for Bmh2-GFP (Figure S3, left).We next asked if 14-3-3 localization is affected in an sps1D background.We see both nuclear and cytoplasmic localization in the absence of SPS1 (Figure 5C, bottom and Figure S3, middle), suggesting that SPS1 is not required for proper 14-3-3 localization.Finally, we asked if 14-3-3 localization was dependent on the presence of both isoforms.To answer this question, we examined BMH1-GFP in a bmh2D background and we saw no obvious difference in localization in sporulating cells (Figures S3 and S4).
Yeast deficient in 14-3-3 isoforms exhibit reduced sporulation efficiency Because we see a defect in sporulation efficiency in cells carrying the sps1-T12A allele, we asked if 14-3-3s play a role in sporulating cells.We examined strains that lack either BMH1 or BMH2 (LH980: bmh1D/bmh1D and LH981: bmh2D/bmh2D) and saw that both had slight reductions in sporulation efficiency (Table 3).Cells lacking BMH1 form spores at 71.0%, while cells lacking BMH2 form spores 70.8% of the time, compared to a wild type value of 80.6% (P,0.02 and P,0.03 respectively compared to wild type, Student's t test).To examine whether the slight sporulation efficiency defect in each bmh mutant was due to the missing 14-3-3 isoform, or the reduction in the total number of 14-3-3 genes, we examined sporulation in the doubly heterozygous LH979 strain (bmh1D/+ bmh2D/+), which sporulated at 79.6%, similar to wildtype (Table 3).
We also examined the effects of further reductions in 14-3-3 gene dosage.Although a strain lacking both bmh1D and bmh2D was viable in the SK1 strain background, this strain displayed severe growth defects and was unable to enter meiosis (likely because of an inability to successfully grow in YPA pre-sporulation medium, as 14-3-3s have been shown to be required for growth in media containing acetate [33]), precluding our ability to assay the sporulation efficiency of the bmh1D bmh2D double mutant strain.However, we were able to examine sporulation in LH982 (bmh1D/ bmh1D bmh2D/+) and LH983 (bmh1D/+ bmh2D/bmh2D), and found that the spore efficiency defect became more severe as the gene dosage of 14-3-3 was further reduced.Specifically, we see that bmh1D/bmh1D bmh2D/+ and bmh1D/+ bmh2D/bmh2D produced 55.8% and 50.0%spores respectively (both P,0.01 compared to wild type, Student's t test).Taken together, these results suggest a role for 14-3-3 proteins in sporulation efficiency (Table 3).

Localization of Sps1 and Sps1-T12A during sporulation
Having shown that Sps1, Bmh1 and Bmh2 are all expressed at the same time during sporulation and that they are capable of physically interacting, we investigated if Sps1 localization resembled that of the 14-3-3s.While a previous report suggested an N- terminal GFP-Sps1 fusion contained within strain Y5050 [11] localizes to the prospore membrane, we were unable to detect such localization.Amplification and sequencing of the GFP-SPS1 locus from Y5050 revealed that the start codon (ATG) was absent from the construct, indicating that it does not produce the intended protein product (Figure S5).
To determine Sps1 localization, we constructed an N-terminal sfGFP-SPS1 fusion by integrating a monomeric variant of the fast-folding superfolderGFP (sfGFP) [67][68][69] at the SPS1 locus.We see sfGFP-Sps1 in both the cytoplasm and the nucleus (Fig. 6 top).As spores matured and became refractile, nuclear localization became more distinct and cytoplasmic localization was reduced, compared to cells earlier in the sporulation process (Fig. 6 top).As 14-3-3 proteins are known to affect the localization of their target proteins, we wanted to determine if mutation of threonine 12 to alanine affected Sps1 localization.Using sfGFP-sps1-T12A, we saw a similar pattern of localization as we did in sfGFP-SPS1 (Figure 6 bottom), suggesting that 14-3-3 proteins do not affect Sps1 localization.
We then asked whether this sequence was sufficient to drive the nuclear localization of a protein that is normally not present in the nucleus.A GFP-GST fusion was created, in which the addition of GST prevents the passive diffusion into the nucleus of GFP [71].We observed an increase in nuclear localization of the fusion protein containing an intact NLS in comparison to those with just the GFP-GST fusion or a GFP-GST fusion with mutated NLS (Fig. 7C).Therefore, Sps1 has a functioning NLS that is both necessary and sufficient for nuclear localization.

Discussion
Sps1 is a serine/threonine kinase that is expressed during sporulation and functions in the process of spore formation.Interestingly, genome-wide studies have identified targets of yeast 14-3-3 proteins [16,72], although none of these studies have identified Sps1 as a target for either Bmh1 or Bmh2.
In this study, we use phylogeny to show that Sps1 is a bona-fide GCKIII kinase, and identify three regions important for its function: a conserved C-terminal motif containing the invariant ExxxPG at residues 464-469, a phosphorylation site T12, and a nuclear localization signal from 411 to 415 (KKHKK).We also show that T12 is part of a 14-3-3 binding site, and is required for the physical interaction between Sps1 and the 14-3-3 isoforms Bmh1 and Bmh2.We demonstrate that 14-3-3s are important for sporulation, as cells lacking in BMH1 and BMH2 have defects in sporulation efficiency and show a genetic interaction with SPS1.Because we see both a genetic and physical interaction, we propose that SPS1 and the 14-3-3s act together to regulate sporulation.

Sps1 is a GCKIII family member
While an initial analysis suggested Sps1 was an outlier with respect to the Ste20 family [5], our bioinformatic analyses indicate that Sps1 belongs to the GCKIII family of Ste20 kinases.The identification of Sps1 as a GCKIII emerged using two different phylogenetic reconstruction strategies, supporting the robustness of the determination.The phylogenetic results also indicate that Sps1 is present in a distinct clade of animal and plant GCKIII kinases, indicating that the GCKIII kinases are an ancient and distinct subfamily of Ste20 kinases, likely present in the common eukaryotic ancestor ,1 billion years ago.
Our comparison of Sps1 with other GCKIII kinases led to the identification of a previously unappreciated conserved region found in all GCKIII kinases.This region is found at the Cterminus of the regulatory domain, and we call this region the EPG motif.Our experimental results demonstrate the importance of this region and suggest that the EPG motif may play an important role in other GCKIII kinases.
We also identify a nuclear localization sequence on Sps1, which is both necessary and sufficient for directing nuclear localization (residues 411-415).Other GCKIII proteins have also been reported to localize to both the nucleus and cytoplasm, and the nuclear localization domain of Mst3 has been mapped to residues 278-294 [73]; the location of this nuclear localization signal is conserved in the mammalian GCKIIIs and in Drosophila, but has changed in Sps1.

Role of BMH1 and BMH2 in sporulation
Here we show that BMH1 and BMH2 are important for the efficient formation of spores in the SK1 background.Previous studies have not identified a role in sporulation for BMH1, including a genome-wide study using the yeast deletion collection in the S288c background [74].It is possible that loss of BMH1 and BMH2 does not have a sporulation defect in the S288c background, or it could be that the less efficiently sporulating S288c strain background made it more difficult to detect the mild sporulation defect we see in bmh1 and bmh2 mutants.
In most yeast strain backgrounds examined, the bmh1 bmh2 double mutant is inviable.In the SK1 strain, the bmh1 bmh2 mutant is viable, but grows very slowly and produces cells of abnormal morphology, precluding the ability to examine the double mutant during sporulation.Because of this, we examined sporulation in a bmh1/+ bmh2/bmh2 strain and see that this strain sporulates at 50% (compared to wild type levels of about 81%), a more severe phenotype than that seen with either single mutant.Our results suggest that if we were able to examine the sporulation defect in a bmh1 bmh2 double mutant, we may see a more severe sporulation efficiency of less than 50%.
Why are there two 14-3-3 isoforms in yeast?In higher eukaryotes, there are several isoforms of 14-3-3 proteins, and these different isoforms are hypothesized to play different roles [28,75].The yeast Bmh1 and Bmh2 are considered paralogs, and likely arose during a whole genome duplication event [76].The yeast 14-3-3 isoforms appear to be largely redundant in terms of function, although specific phenotypes have been associated with the loss of only a single isoform.For instance, the loss of BMH1 causes an increase in glycogen accumulation [77] whereas loss of BMH2 results in abnormal accumulation of polyphosphate [78].
We see that 14-3-3 isoform levels shift during sporulation.In vegetatively growing cells, we see that Bmh1 is more prevalent compared to Bmh2, as previously described [65].However, during sporulation, Bmh2 levels rise in comparison to Bmh1 levels.We do not know whether this relative increase in Bmh2 level is important for 14-3-3 function or is merely a consequence of transcriptional regulation changes that occur during sporulation.

14-3-3 regulation of Sps1 during sporulation
Our data suggests that Bmh1 and Bmh2 are important for the positive regulation of Sps1 function during sporulation, since the phenotypes of the strains containing sps1-T12A are similar to those lacking bmh1, and bmh2.As Bmh1 and Bmh2 physically interact with Sps1 in a manner that depends on the T12 within Sps1, these data raise the possibility that this interaction is functionally significant.
We propose that the interaction with 14-3-3s modulates Sps1 function but is not absolutely required for Sps1 activity, because an sps1 null allele has a much more severe phenotype than either the sps1-T12A or the bmh1 or bmh2 mutants.In comparison to cells carrying sps1D allele, which rarely produce refractile spores, cells with the sps1-T12A allele display a less severe phenotype, sometimes producing spores that appear normal with respect to their ability to form the outer layers of the spore wall, to package the appropriate number of spores within an ascus, and to germinate.
Although we cannot directly assay the phenotype of the bmh1 bmh2 double mutant, we anticipate that it would be more severe than the sps1-T12A phenotype because cells carrying sps1-T12A allele sporulate more efficiently than the bmh1/+ bmh2/bmh2 mutant cells (Tables 2 and 3).The more severe defect seen in the bmh1/+ bmh2/bmh2 mutant along with more severe defect seen in bmh1D/+ bmh2D/+ SBP-sps1-T12A/sps1D compared to SBP-sps1-T12A/sps1D (Tables 3 and 4) suggests that Sps1 may not be the only relevant binding partner of 14-3-3 proteins during sporulation.Alternatively, it is possible that T12 is not the only residue on Sps1 that can mediate the Sps1-14-3-3 interaction.It is possible that the weak phenotype seen with sps1-T12A compared to the bmh1/+ bmh2/bmh2 mutant is due to the presence on Sps1 of other 14-3-3 interaction sites.
Other interactions with 14-3-3 proteins and GCKIII kinases have been identified.14-3-3f is a substrate of YSK1/SOK1, and its phosphorylation has been shown to be important for Golgi positioning [81] and to disrupt the binding of the proapoptotic factor ASK to14-3-3f during apoptosis [82].
Interestingly, there may be a role for 14-3-3s in modulating Ste20 family kinases in general.In the S1278 background, 14-3-3s have been found to bind directly to the C-terminal kinasecontaining portion (from amino acid 494 to 939) of the Ste20 kinase, a PAK family member [37].Since the domain architecture of PAKs and GCKs are reversed, and since Bmh1 and Bmh2 interacts with the C-terminal half of Ste20 and the N-terminal half of Sps1 (which are both STE20 family kinases), it is tempting to speculate that Bmh1 and Bmh2 may modulate pseudohyphaldevelopment directly through the modulation of Ste20 activity, especially because our analysis of amino acids 494-939 of Ste20 using Scansite predicts a 14-3-3 binding site of medium stringency surrounding threonine 546, which has been found to be phosphorylated by a number of different studies [83][84][85].

Figure 1 .
Figure 1.Sps1 is a GCKIII family member.Phylogenetic analysis using maximum likelihood-based methods.Branch support Ps are noted where relevant.(A.t.: Arabidopsis thaliana, D.m.: Drosophila melanogaster, H.s.: Homo sapiens, S.c.: Saccharomyces cerevisiae.)The GCK subfamilies, the PAK subfamily, and the MEKK outgroup are highlighted in different colors.The S. cerevisiae protein names are in blue; all other kinase names are in black.doi:10.1371/journal.pone.0113528.g001

Table 1 .
The EPG motif containing region is required for SPS1 function.

Table 3 .
BMH1 and BMH2 are required for sporulation.