Phosphoregulation of the cytokinetic protein Fic1 contributes to fission yeast growth polarity establishment

Polarization is a common feature of eukaryotic and prokaryotic cells (Hu and Lutkenhaus, 1999; Miller and Johnson, 1994). Multicellular organisms couple polarization events in neighboring cells to drive key developmental processes (Moorhouse et al., 2015). In a single cell, polarization governs processes such as growth, motility and fate specification (Mortimer et al., 2008; Pham et al., 2015; Matsuoka and Masahiro, 2018).

The fission yeast Schizosaccharomyces pombe is a powerful model organism for studying mechanisms by which polarization is established, maintained and modified (Arellano et al., 1999; Miller and Johnson, 1994; Ottilie et al., 1995). S. pombe is a rod-shaped organism, with growth limited to its cell tips (Streiblova and Wolf, 1972). After cell division, elongation occurs first only at old ends inherited from mother cells. Then, at a later point, known as new end take off (NETO), cells transition to bipolar growth by also extending at new ends established by the most recent cell division (Mitchison and Nurse, 1985).

Historically, NETO was thought to be triggered when cells reached a minimal cell size and completed S-phase (Mitchison and Nurse, 1985). However, NETO also requires proper completion of cytokinesis (Bohnert and Gould, 2012). Specifically, loss of the contractile ring (CR) protein Fic1 (Roberts-Galbraith et al., 2009) leads to abnormal persistence of CR components at new ends and curbs NETO even if factors responsible for growth are properly positioned at new cell ends (Bohnert and Gould, 2012). Barriers to NETO, caused by loss of Fic1 or other late cytokinetic factors, in turn promote growth orientations that favor a dimorphic switch from a unicellular state to a more-invasive, pseudohyphal form (Bohnert and Gould, 2012).

Our growing knowledge of NETO highlights the role of protein kinases at cell tips (Arellano et al., 2002; Fujita and Misumi, 2009; Grallert et al., 2013; Kettenbach et al., 2015; Kim et al., 2003; Kume et al., 2017, 2011; Martin et al., 2005). Polarity kinases also target CR proteins and influence their localization and cytokinetic function (Bhattacharjee et al., 2020; Lee et al., 2018; Magliozzi et al., 2020). The interplay between the fidelity of cytokinesis and proper polarity establishment in the next cell cycle, demonstrated by the role of Fic1, may therefore be phosphoregulated.

Here, we show that Fic1 is phosphorylated at two C-terminal residues. Although Cdk1 and the casein kinase II Orb5 each phosphorylate one of the sites in vitro, we found that none of the 111 S. pombe protein kinases are solely responsible for phosphorylation of either site. Fic1 phospho-mimetic and phospho-ablating mutations impaired S. pombe NETO and produced an invasive pseudohyphal phenotype, indicating phosphorylation controls its role in polarity. Our findings predict complex regulation of cytokinesis-based polarity determinants and suggest at least two different groups of kinases influence polarity by modulating the Fic1 phosphostate.

As shown previously (Bohnert and Gould, 2012), immunoblotting of Fic1–FLAG₃ immunoprecipitates revealed four distinct bands from untreated cells, but only one band when immunoprecipitates were treated with phosphatase (Fig. 1A). Thus, Fic1 is a phosphoprotein, and, given the multiple Fic1–FLAG₃ species, we conclude that Fic1 is phosphorylated on multiple residues.

To assess whether Fic1 phosphostatus is cell cycle regulated, we analyzed Fic1–FLAG₃ gel mobility after cell cycle arrests at different points, either through the use of temperature-sensitive alleles (cdc10-V50 G1 arrest, cdc25-22 G2 arrest, nda3-KM311 prometaphase arrest, or cps1-191 cytokinesis arrest) or after addition of hydroxyurea (S-phase arrest). In all cases, Fic1–FLAG₃ gel mobilities were identical (Fig. 1B). We corroborated this result using a cdc25-22 block-and-release experiment, in which samples were taken following release from a G2 arrest. Fic1–FLAG₃ gel mobility shifts were identical at each time point (Fig. 1C), verifying that multiple phospho-species of Fic1 exist throughout the cell cycle.

The C-terminus of Fic1 (amino acids 127–end, ‘Fic1C’; Fig. 2A) localizes to the CR but not to cell tips, and this fragment is necessary and sufficient for NETO (Bohnert and Gould, 2012). In contrast, the N-terminal C2 domain (amino acids 1–126, ‘Fic1N’; Fig. 2A) neither localizes to the CR nor contributes to bipolar growth establishment, but is required for anchoring Fic1 to cell tips (Bohnert and Gould, 2012). Whereas Fic1N–GFP did not migrate as multiple species on SDS-PAGE gels (Fig. 2B), Fic1C–GFP showed a phosphoshift that was abrogated by phosphatase treatment (Fig. 2C). These results suggested that Fic1 phosphorylation might affect its function at the CR and, therefore, in NETO.

If this were the case, we expected that Fic1 phosphorylation would occur even if Fic1 lost its ability to anchor at cell tips. Based on homology to S. cerevisiae Inn1 (Devrekanli et al., 2012; Sanchez-Diaz et al., 2008), we predicted that two lysine residues within the C2 domain (Fig. 2A) mediate cell-tip localization, and we mutated them to alanine residues. Cell tip localization of Fic1-K22A,K27A–GFP was greatly reduced compared to that of Fic1–GFP (Fig. 2D,E). Additionally, whereas wild-type Fic1-GFP localized broadly across cell tips, Fic1-K22A,K27A–GFP localization at cell tips was restricted to puncta that also contained Cdc15–mCherry, a tip protein and Fic1 interactor (Roberts-Galbraith et al., 2009) (Fig. 2D,E). However, Fic1-K22A,K27A promoted proper bipolar growth (Fig. 2F–H), targeted to the CR (Fig. 2I), and was phosphorylated to the same extent as wild-type Fic1 (Fig. 2J), consistent with the idea that Fic1 phosphorylation influences its function at the CR to modulate polarity.

Fic1 phosphorylation sites have not been identified in proteome-wide screens (Lock et al., 2019). Thus, we used mass spectrometry of tandem affinity-purified Fic1–TAP to identify phosphorylation sites in a targeted manner. Phosphorylation of two C-terminal residues, T178 and S241, was identified (Fig. 3A,B; Fig. S1A). T178 and S241 were each mutated to alanine to abolish phosphorylation, or to aspartate to potentially mimic constitutive phosphorylation. These phosphomutants were then integrated at the endogenous fic1 locus, tagged with FLAG₃, and tested for alteration in SDS-PAGE mobility. Alanine mutations of T178 or S241 individually eliminated two of the four bands, indicating that one band, the upper band, represented dually phosphorylated protein and each intermediate band represented singly phosphorylated Fic1 (Fig. 3C), also confirming that these two residues are the major Fic1 phosphosites. Consistent with this interpretation, Fic1-T178A,S241A (Fic1-2A) migrated as a single band (Fig. 3C). In Fic1 aspartate mutants, similar gel mobility patterns were observed, except that all bands were slightly retarded in mobility (Fig. 3C). Thus, phosphorylation occurs individually and in combination at T178 and S241 in vivo.

Because phosphorylation occurs in Fic1C, which is important for polarity establishment, we set out to identify the kinase(s) responsible. T178 and S241 fit the consensus sequences for CDK (S/T-P) and CK2 (S-X-X-E/D), respectively (Fig. 3B) (Meggio et al., 1994; Nigg, 1993), and these kinases are important for polarized growth (Adams et al., 1990; McCusker et al., 2007; Rethinaswamy et al., 1998; Shimada et al., 2000; Snell and Nurse, 1994). In vitro kinase assays using Cdk1–cyclinB, S. pombe CK2 (Orb5) or human CK2 with His₆–Fic1, His₆–Fic1-T178A, and/or His₆–Fic1-S241A revealed that they can phosphorylate the sites that fit their consensus sequence (Fig. 3D,E). His₆–Fic1-T178A phosphorylation by Cdk1 was significantly reduced compared to that of His₆–Fic1, demonstrating that in vitro Cdk1 primarily targets T178 (Fig. 3D). To test whether Cdk1 and Orb5 phosphorylate Fic1 in vivo, we assayed the Fic1 phosphostatus in analog-sensitive and temperature-sensitive mutants of these kinases. The phosphostate of Fic1 was unaltered in these single and double mutants (Fig. 3F). Thus, Cdk1 and Orb5 are not solely responsible for Fic1 phosphorylation in vivo.

We next took an unbiased approach to try to identify the kinase(s) responsible for Fic1 phosphorylation. Starting with polarity kinases, then extending to all kinases, we screened gene deletions of individual non-essential kinases, temperature or analog-sensitive mutants of essential protein kinase genes, and combination mutants of paralogs (e.g. pck1-as pck2-as) for changes to the phosphostate of Fic1 (Bimbó et al., 2005; Chen et al., 2014; Cipak et al., 2011; Gregan et al., 2007; Kim et al., 2010). In the course of this screening, we determined that several kinase deletion strains in Bioneer V3 (Kim et al., 2010) contained not only a targeted deletion allele but also the wild-type kinase gene. Thus, we constructed new deletion mutants of atg1, hal4, lsk1, mak2, mek1, sty1, ppk24, ppk34 and wis1. Through immunoblotting, we did not detect loss of either Fic1 phosphorylation event in any of the 111 single or 11 combination kinase mutants tested (Figs S1B and S2), indicating that multiple kinases can phosphorylate Fic1.

We next tried to identify pathways regulating the Fic1 phosphostate. To this end, we treated cells with osmotic, oxidative, cell wall integrity, and plasma membrane stressors (Cadou et al., 2010; Chen et al., 2008; Madrid et al., 2006; Robertson and Hagan, 2008). None of the stressors tested evoked a change in the phosphostate of Fic1 (Fig. 3G). However, phosphorylation at S241, but not T178, was lost after treatment with a high dose of 1-NM-PP1 (Fig. 3H). 1-NM-PP1 is an inhibitor designed to preferentially target kinases with space-creating mutations in their ATP-binding pocket. However, some kinase families such as Src, CDK and CAMKII, are sensitive to high levels of 1-NM-PP1 (Bishop et al., 2000). Loss of phosphorylation on S241 but not on T178 suggests that distinct groups of kinases phosphorylate each site, one of which can be inhibited by high levels of 1-NM-PP1. Collectively, these data establish that Fic1 phosphoregulation involves multiple kinases that are apparently coordinated to keep the ratios of Fic1 phosphorylation events similar throughout the cell cycle and under different physiological conditions.

To assess whether disrupting the phosphorylation state of Fic1 altered its localization, we analyzed the fluorescence intensities of mNeonGreen-tagged Fic1 variants at the CR and cell tips. We observed no differences between Fic1 and the Fic1 phosphomutants (Fig. 4A,B). Next, we performed time-lapse imaging to determine whether Fic1 CR recruitment timing relative to spindle pole body (SPB) separation differed for either phosphomutant. We found that Fic1, Fic1-2A, and Fic1-2D were recruited to the CR with similar timing (Fig. 4C).

Fic1 localizes to the CR via its interaction with the SH3 domains of the F-BAR proteins, Cdc15 and Imp2 (Roberts-Galbraith et al., 2009; Ren et al., 2015). Because one Fic1 phosphosite (S241) is proximal to the P254-P257 PxxP motif required for these SH3 domain interactions (Bohnert and Gould, 2012), we tested whether phosphorylation interferes with Cdc15 and Imp2 binding. Consistent with normal CR recruitment, both Fic1 phosphomutants co-immunoprecipitated with Cdc15 and Imp2 (Fig. S3A), indicating that deregulation of Fic1 phosphorylation does not grossly alter these known interactions.

To assess the relevance of Fic1 phosphorylation to NETO, we analyzed the growth polarity of the phosphomutants (Fig. 4D). fic1-2A and fic1-2D had similar levels of monopolar cells to those in fic1Δ, whereas each individual phosphomutant displayed levels that were intermediate between wild type and fic1Δ (Fig. 4D–F). Also, as expected given that wild-type cells commonly initiate NETO by late interphase, nearly all fic1⁺ cells arrested in late G2 exhibited bipolar growth (Fig. S3B,C). In contrast, a high percentage of cdc25-22 fic1-2A and cdc25-22 fic1-2D cells were still monopolar, like cdc25-22 fic1Δ (Fig. S3B,C). By using time-lapse DIC imaging, we confirmed that the polarized growth defects in fic1Δ, fic1-2A and fic1-2D were specific to new ends (Fig. 4G; Fig. S3D). Although the phosphomutants did not support proper NETO, they retained some function because they are not synthetically lethal with pxl1Δ, ppb1Δ or sid2-250 like fic1Δ (Bohnert and Gould, 2012) (Fig. S3E), possibly because they can still associate with Cdc15 and Imp2 (Fig. S3A). Considering our results together, we hypothesize that phosphorylation affects the interaction of Fic1 with unknown factor(s) at the CR that influences its CR function.

Inability to support proper NETO is exhibited by S. pombe and S. japonicus cells that have undergone the dimorphic switch from single-celled to pseudohyphal growth (Dodgson et al., 2010; Sipiczki et al., 1998), and fic1Δ cells show increased invasive pseudoyhyphal growth compared to that shown by wild-type cells (Fig. 4H,I; Fig. S3F) (Bohnert and Gould, 2012). Consistent with fic1 phosphomutants possessing growth polarity defects akin to fic1Δ, fic1-2A and fic1-2D cells formed pseudohyphal structures invading the agar (Fig. 4H; Fig. S3F,G). Also, each individual Fic1 aspartate mutant was more invasive than wild-type (Fig. S3F). Thus, the dimorphic switch from single-celled to pseudohyphal form may involve Fic1 phosphoregulation.

In conclusion, although phosphorylation serves diverse roles during eukaryotic cytokinesis (Bohnert and Gould, 2011), it has been unclear whether CR protein phosphorylation impacts cellular processes other than cell division. In this study, we found that Fic1 phosphorylation influences polarity and the transition to hyphal growth. Our findings support the ideas that (1) regulating CR function can directly impact the dimorphic switch; and (2) modulation of kinase and/or phosphatase signaling may be sufficient for this switch. As extensive phosphosignaling occurs during hyphal growth (Sudbery, 2011), integration of multiple cues likely guarantees the robustness of this transition.

S. pombe strains (Table S1) were grown in yeast extract with supplements (YES) or Edinburgh minimal medium with relevant supplements. Genes were tagged at the 3′ end of their ORFs with sequences encoding GFP:kan^R, HA₃:hyg^R, FLAG₃:kan^R, mNeonGreen:kanR or FLAG₃:hyg^R using pFA6 cassettes as previously described (Bähler et al., 1998; Wach et al., 1994). A lithium acetate method (Keeney and Boeke, 1994) was used in S. pombe tagging transformations, and integration of tags was verified using whole-cell PCR and/or microscopy. Introduction of tagged loci into other genetic backgrounds was accomplished using standard S. pombe mating, sporulation and tetrad dissection techniques. For arresting cdc25-22 and cps1-191, cells were grown at 25°C and then shifted to 36°C for 3 h. nda3-KM311 arrest was achieved by growing cells at 32°C and then shifting to 18°C for 6.5 h. For blocking of cdc10-V50, cells were grown at 25°C and then shifted to 36°C for 4 h. S-phase arrest was achieved by treating cells with 12 mM hydroxyurea for 4 h at 32°C.

Mutants and truncations of fic1 were expressed from the endogenous fic1⁺ locus. To make these strains, a pIRT2 vector was used in which fic1⁺ gDNA with 5′ and 3′ flanks was inserted between BamHI and PstI sites of pIRT2 (Bohnert and Gould, 2012). Mutations were then introduced via site-directed mutagenesis and confirmed by DNA sequencing. fic1Δ was transformed with these pIRT2-fic1 constructs, and stable integrants resistant to 1.5 g/l 5-fluoroorotic acid (5-FOA) were isolated and confirmed by whole-cell PCR, DNA sequencing and immunoblotting.

To construct analog-sensitive protein kinase strains, the coding sequences with 5′ and 3′ flanks of pck1⁺, pck2⁺ and nnk1⁺ were PCR amplified from S. pombe genomic DNA using PrimeSTAR GXL DNA polymerase (Takara) and ligated into the pCR-Blunt II-TOPO^® vector (Invitrogen). The resulting inserts were verified by sequencing. The gate-keeper residues in Pck1 and Pck2 kinases were identified as M744 and M763, respectively, and in Nnk1 kinase as M537 (Gregan et al., 2007). These residues were mutated to glycine or alanine using mutagenic oligonucleotide primers and QuikChange II site-directed mutagenesis kit (Stratagene). The desired mutations were verified by sequencing. Next the pck1- and pck2-containing plasmids were linearized and transformed into pck1::ura4⁺ and pck1::ura4⁺ cells, respectively, and the plasmids containing nnk1-as mutations were transformed into nnk1::ura4⁺-HA₃-TAP:kan^R P_nmt41-GBP-mCherry-nnk1-leu1⁺, using a lithium acetate method (Keeney and Boeke, 1994). Transformants were selected based on resistance to 5-FOA and then confirmed first by colony PCR and then by DNA sequencing. Next, the allele P_nmt41-GBP-mCherry-nnk1-leu1⁺ was crossed out to obtain nnk1(M537G)-HA₃-TAP:kan^R and nnk1(M537A)-HA₃-TAP:kan^R.

For serial-dilution growth assays, cells were grown in liquid YE at 32°C; then, three serial 1:10 dilutions starting at 4×10⁶ were created, 2 µl of each dilution was spotted on YE agar and cells were grown at the indicated temperatures for 3–5 days.

Cells were lysed by bead disruption in NP40 lysis buffer in denaturing conditions as previously described (Gould et al., 1991), except with the addition of 0.5 mM diisopropyl fluorophosphate (Sigma-Aldrich). Immunoblot analysis of cell lysates and immunoprecipitates was performed using anti-FLAG (M2; Sigma-Aldrich; 1:5000; cat. no. F1804) or anti-PSTAIRE Cdc2 (Sigma-Aldrich; 1:5000; cat. no. P7962) antibodies or serums raised against GST-Cdc15 (VU326; 1:2000) (Roberts-Galbraith et al., 2009) or His₆-Imp2 (VU483; 1:1000) (McDonald et al., 2016) as previously described (Bohnert et al., 2009). For gel shifts, denatured samples were treated with lambda-phosphatase (New England Biolabs) in 25 mM HEPES-NaOH (pH 7.4), 150 mM NaCl, and 1 mM MnCl₂ and incubated for 30 min at 30°C with shaking.

In vitro kinase assays with kinase-active Cdk1 were performed as described by Yoon et al. (2006) using a kinase buffer consisting of 50 mM Tris-HCl pH 7.4, 10 mM MgCl₂ and 2 mM DTT supplemented with 10 µM cold ATP and 5 µCi γ-[³²P]ATP. Reactions contained 100 ng of kinase-active Cdk1 and 1 µg of recombinant His₆–Fic1, His₆–Fic1–T178A or His₆–Fic1-S241A. The addition of sample buffer and boiling terminated the reactions. Samples of each reaction were separated by SDS-PAGE and visualized by Coomassie Blue staining and autoradiography.

Tandem affinity purifications (TAPs) for Fic1–TAP proteins were conducted as described previously (Gould et al., 2004). Proteins were subjected to mass spectrometric analysis on an LTQ Velos by 3-phase multidimensional protein identification technology, as previously described (McDonald et al., 2002; Chen et. al., 2013) with modifications. Proteins were resuspended in 8 M urea buffer (8 M urea in 100 mM Tris-HCl, pH 8.5), reduced with Tris (2-carboxyethyl) phosphine, alkylated with 2-chloro acetamide, and digested with trypsin or chymotrypsin. The resulting peptides were desalted by C-18 spin column (Pierce). Raw mass spectrometry data were filtered with Scansifter and searched by SEQUEST algorithm. Scaffold (version 4.4.8 or version 4.2.1) and Scaffold PTM (version 3.0.1) (both from Proteome Software, Portland, OR) were used for data assembly and filtering. Phosphorylation sites were filtered to 50% Scaffold localization using Scaffold PTM (v2.0, Proteome Software, Portland, OR) and Ascores (Beausoleil et al., 2006).

Live-cell images of S. pombe were acquired using one of the following: (1) a spinning disc confocal microscope (Ultraview LCI, PerkinElmer) equipped with a 100× NA 1.40 PlanApo oil immersion objective, a 488-nm argon ion laser (GFP and mNeonGreen), a 594-nm helium neon laser (mCherry), a charge-coupled device camera (Orca-ER, Hamamatsu Phototonics), and Metamorph 7.1 software (MDS Analytical Technologies and Molecular Devices) or (2) a personal DeltaVision microscope system (Applied Precision) that includes an Olympus IX71 microscope, 60× NA 1.42 PlanApo and 100× NA 1.40 UPlanSApo objectives, a Photometrics CoolSnap HQ2 camera, and softWoRx imaging software. All cells were in log phase growth before temperature-sensitive shifts and/or live imaging.

For Calcofluor staining, cells were washed in PBS and then resuspended in PBS containing 5 µg/ml Calcofluor. After incubation on ice for 30 min, cells were washed three times in PBS and images were acquired. Using the proximity of birth scars to cell ends, growth/morphology was scored as one of the following: monopolar (i.e. growth on one end), bipolar (i.e. growth on both ends), monopolar and septated, bipolar and septated, or multiseptated. For cells just completing division, daughter cells were scored as monopolar as long as ingression of the mother cell had progressed to such a degree that birth scars could be easily identified at new ends. All cells stained with Calcofluor were grown to log phase at 25°C, except that cdc25-22 mutants were grown overnight at 25°C and then shifted to 36°C for 3 h before staining.

Intensity measurements were made with ImageJ software (Schindelin et al., 2012) using non-deconvolved summed Z-projections of the images. For all intensity measurements, the background was subtracted by creating a region of interest (ROI) in the same image in an area clear of cells. The background raw intensity was divided by the area of the background, which was multiplied by the area of the ROI. This number was subtracted from the raw integrated intensity of that ROI. To account for autofluorescence, cells lacking fluorescent tags but otherwise of isogenic backgrounds, were imaged and the fluorescence intensity per pixel was quantified from summed Z-projections of the images by subtracting the background intensity from the measured raw integrated intensity of the ROI before dividing the raw integrated intensity of the ROI by the area of that ROI. This autofluorescence per pixel measurement was multiplied by the area of the ROI from fluorescent cells before subtracting this value from the from the raw integrated intensity of that ROI. Representative images are max intensity Z-projections.

Time-lapse imaging was performed using an ONIX microfluidics perfusion system (CellASIC ONIX; EMD Millipore). A suspension of 50 µl of 40×10⁶ cells/ml YE was loaded into Y04C plates for 5 s at 8 psi. YE medium was flowed through the chamber at 5 psi throughout imaging.

Images of yeast cells and pseudohyphae on YE agar plates were acquired by focusing a camera (PowerShot SD750; Canon) through a microscope (Universal; Carl Zeiss) equipped with a 20× NA 0.32 objective.

To assay pseudohyphal invasion into 2% agar, 5 μl containing a total of 10⁵ cells were spotted on 2% YE agar and incubated at 29°C for 20 days. Colonies were subsequently placed under a steady stream of water, and surface growth was wiped off using a paper towel, as described previously (Pöhlmann and Fleig, 2010; Prevorovsky et al., 2009).

We thank Alaina Willet, Sierra Cullati, and Chloe Snider for critical reading of the manuscript and Malwina Huzarska for technical assistance. Some data in the paper formed part of Kenneth Adam Bohnert's PhD thesis in the Department of Cell and Developmental Biology at Vanderbilt University in 2013.