NDR kinase Sid2 drives anillin-like Mid1 from the membrane to promote cytokinesis and medial division site placement

SUMMARY

In animals and fungi, cytokinesis is facilitated by the constriction of an actomyosin contractile ring (CR) [1]. In Schizosaccharomyces pombe, the CR forms mid-cell during mitosis from clusters of proteins at the medial cell cortex called nodes [2]. The anillin-like protein Mid1 localizes to nodes and is required for CR assembly at mid-cell [3]. When CR constriction begins, Mid1 leaves the division site. How Mid1 disassociates and whether this step is important for cytokinetic progression has been unknown. The Septation Initiation Network (SIN), analogous to the Hippo pathway of multicellular organisms, is a signaling cascade that triggers node dispersal, CR assembly and constriction, and septum formation [4,5]. We report that the terminal SIN kinase, Sid2 [6], phosphorylates Mid1 to drive its removal from the cortex at CR constriction onset. A Mid1 mutant that cannot be phosphorylated by Sid2 remains cortical during cytokinesis, over-accumulates in interphase nodes following cell division in a manner dependent on the SAD kinase Cdr2, advances the G2/M transition, precociously recruits other CR components to nodes, pulls Cdr2 aberrantly into the CR, and reduces rates of CR maturation and constriction. When combined with cdr2 mutants that affect node assembly or disassembly, gross defects in division site positioning result. Our findings identify Mid1 as a key Sid2 substrate for SIN-mediated remodeling of the division site for efficient cytokinesis and provide evidence that nodes serve to integrate signals coordinating cell cycle progression and cytokinesis.

eTOC blurb:

Willet et al. identify anillin-like Mid1 as a substrate of the NDR-related SIN kinase, Sid2. Sid2 phosphorylation of Mid1 disrupts Mid1 interaction with membrane and SAD kinase Cdr2, and contributes to SIN-mediated remodeling of the fission yeast division site for efficient cytokinesis and accurate division site placement in the next cell cycle.

Graphical Abstract

RESULTS and DISCUSSION

NDR kinase Sid2 phosphorylates Mid1 in vitro and in vivo

During interphase, S. pombe Mid1 shuttles between the nucleus and Type1 cortical nodes established by Cdr2, concentrating at nodes in late interphase [7–9]. Type2 cortical nodes, consisting of RhoGEF Gef2, Kinesin-like Klp8, Nod1, and Blt1, emerge from CR remnants and diffuse to the cell equator where they are captured by Type1 nodes at mitotic onset to become cytokinetic nodes [8]. At this time, Gef2 links Mid1 to Blt1 to stabilize Mid1 node association; Cdr2 also stabilizes this complex [10]. Next, Mid1 recruits other cytokinetic node proteins including the myosin II Myo2, myosin light chain Rlc1, IQGAP Rng2, F-BAR Cdc15, and formin Cdc12, and Mid1 is incorporated into the CR along with Type2 node components [9,11–17]. Type1 node components such as Cdr2 are not incorporated into the CR [8]. When constriction begins, Mid1 leaves the CR and re-accumulates in the nucleus, while Type2 node proteins remain in the CR [8]. Meanwhile, Type1 nodes reset to daughter cell middles to ensure medial cell division in the subsequent cell cycle [8,18]. Although Mid1’s localization to Type1 and cytokinetic nodes depends on its ability to bind the plasma membrane and additional protein interactions modulated by Cdk1 and polo-like kinase Plo1 [7,9,10,19–21], it was previously unknown how or why Mid1 leaves the division site at constriction onset.

We hypothesized that Mid1 is a Sid2 substrate involved in SIN-directed cytokinesis for three reasons. First, the SIN kinase Sid2 translocates from spindle pole bodies (SPB)s to the CR [6] concurrent with Mid1’s departure from the division site [21]. Second, Mid1 contains Sid2 consensus phosphorylation sites, RXXpS [22,23] (Figure 1A), and third, precocious SIN activation disperses Cdr2 and Mid1 from the cell cortex [4]. To test if Mid1 is a direct Sid2 substrate, we incubated immunopurified Sid2-myc₁₃ (obtained from cdc16–116 arrested SIN hyperactive mutant cells) with γ−³²P-ATP and a fragment of Mid1 that contains all of its RXXS motifs. Sid2 phosphorylated this fragment but not when all nine serines within RXXS motifs were mutated to alanine, indicating that Mid1 is a Sid2 substrate (Figure 1B). To determine if these sites are also phosphorylated in vivo, we used mass spectrometry to identify phosphorylation sites in purified Mid1-TAP. Of the nine potential Sid2 sites, we identified seven (Figure S1), five of which had also been identified in other phosphoproteomic screens, and an eighth was identified in another screen [24,25,26]. We integrated a mid1–9A allele, in which these phosphosites had been substituted with alanines, at the endogenous mid1 locus and C-terminally tagged it with myc₁₃. Immunoprecipitated Mid1-myc₁₃ migrated as a broad band on SDS-PAGE, consistent with the existence of multiple phosphoisoforms, that collapsed to a single band upon phosphatase treatment (Figure 1C). The phosphoisoforms of Mid1–9A-myc₁₃ reproducibly migrated faster than those of Mid1-myc₁₃, consistent with the removal of in vivo phosphosites (Figure 1C). Because phosphatase treatment affected the gel mobility of Mid1–9A, these results indicate that its ability to serve as a substrate for protein kinases other than Sid2 remains intact.

Figure 1. NDR kinase Sid2 phosphorylates Mid1 in vitro and in vivo.

(A) Schematic, drawn to scale, of S. pombe Mid1 domain organization. The two nuclear export sequences (NES), one nuclear localization sequence (NLS), one nuclear targeting sequence, the C2 domain, pleckstrin homology (PH) domain, protein binding regions that overlap with phosphorylation sites, the nine identified RXXS consensus motifs for Sid2 and the proposed amphipathic helix are indicated. (B) In vitro kinase assay using His₆-Mid1(1–578) or His₆-Mid1(1–578–9A) incubated with Sid2-myc₁₃ immunoprecipitated from cdc16–116 cells. An anti-myc immunoprecipitate from cdc16–116 sid2⁺ was used as a negative control. Phosphorylation was detected by autoradiography. Loading controls for total Mid1 are shown by Coomassie Blue (CB) staining. (C) Immunoprecipitates from mts3–1 arrested mid1-myc₁₃ or mid1–9A-myc₁₃ cells were treated with phosphatase or buffer control before being resolved by SDS-PAGE and immunoblotted. Brackets span phosphorylated forms. CDK is a loading control. (D) Live-cell images of mid1-mNG and mid1–9A-mNG. Images are max projections. Blue arrows indicate tip-localized Mid1–9A-mNG. Scale bar = 5 μm. (E) Live-cell images of mid1-mNG and mid1–9A-mNG. Images shown are the bottom 3 z-slices max projected and using the Fire look-up table. Scale bar = 2 μm. The number of nodes per cell classified in the yellow, orange or white intensity range were quantified in (F) as described in Methods. Using sum projections of the same z slices, the fluorescence intensity of all nodes per cell was quantified in (G). For (F-G), each quantification was plotted against cell length. (H) mid1-mNG and mid1–9A-mNG in wildtype or cdc16–116 were grown up at 25°C and shifted to 36°C for 1 hou r before imaging. A single medial z-slice is shown. Scale bar = 5 μm. (I) Quantification of the nuclear to cytoplasmic ratio of strains from H using a single medial z-slice. n ≥ 16. Error bars represent SEM. Student’s t-test, **p ≤ 0.01. Also see Figure S1 and Table S1.

Sid2 phosphorylation regulates Mid1’s localization

We next examined the effect of Sid2 phosphorylation on Mid1’s localization. As previously described, Mid1 was detected in the nucleus and cortical nodes surrounding the nucleus in shorter cells, at cytokinetic nodes in longer cells, and at the CR in mitotic cells (Figure 1D) [7,8,17]. Like wildtype, Mid1–9A-mNG localized to the CR in mitotic cells (Figure 1D). Unlike wildtype, Mid1–9A-mNG was difficult to detect in the nucleus and aberrantly localized at cell tips and to bright cortical nodes in both short and long cells (Figure 1D). While the cortical signal of Mid1–9A-mNG appeared brighter than that of Mid1-mNG, the proteins were produced at the same levels as determined by immunoblotting (n = 3, p = 0.23) (Figure 1C).

Because there appeared to be more Mid1–9A-mNG than Mid1-mNG at the medial cortex of cells (Figure 1D), we measured the number of Mid1 nodes and their cortical intensity. Mid1-mNG was present at an increased number of nodes and at increased cortical intensity as cell length increased, and Mid1–9A node number and cortical intensity were increased at all cell lengths compared to wildtype (Figure 1E-G). These data indicate that Sid2 phosphorylation of Mid1 negatively regulates its cortical node localization.

As reported previously [4], Mid1 localization is modulated by SIN activity. In cdc16–116 arrested cells (SIN hyperactive mutant), Mid1 accumulated in the nucleus as indicated by an increased ratio of nuclear to cytoplasmic Mid1 signal (Figure 1H-I). This behavior is consistent with phosphorylation driving Mid1 out of nodes and allowing it to be imported into the nucleus. In contrast, Mid1–9A remained at nodes, rings, and along the septa in cdc16–116 arrested cells; no significant change was observed in its nuclear to cytoplasmic signal (Figure 1H-I). These data are consistent with the idea that Sid2 phosphorylation of Mid1 inhibits its node localization.

To determine if Sid2-mediated phosphorylation affects the kinetics of Mid1 localization, we performed live-cell time-lapse imaging of Mid1-mNG and Mid1–9A-mNG cells. As observed previously [7,8,17,19,21], Mid1-mNG concentrated in nodes in late interphase, then these nodes coalesced into a tight ring in metaphase, and then during CR constriction Mid1 division site intensity waned concurrently with nuclear re-accumulation (Figure 2A-B). In contrast, Mid1–9A-mNG was at nodes throughout interphase, coalesced into a ring during mitosis, and remained at the cell division site during constriction and septation; it also reappeared in faint medial nodes that contained Blt1-mCherry in short cells instead of reaccumulating in the nucleus (Figure 2A-C). Because Mid1–9A localized to the division site after CR constriction onset, we asked if Mid1–9A localized to the CR and/or the membranes lining the septum. Using structured illumination microscopy and a device that positions the cell end-on [27], we found that Mid1–9A-mNG co-localized with CR marker mCherry-Cdc15 in newly formed CRs but not in constricting CRs when it instead lined the septum (Figure 2D). We conclude that Mid1–9A, like Mid1, dissociates from other CR components at CR constriction onset but unlike wildtype, Mid1–9A remains associated with the plasma membrane at the division site. These data further support the conclusion that Sid2-mediated phosphorylation drives Mid1 from the membrane.

Figure 2. Sid2 phosphorylation regulates timing of Mid1 localization.

(A-B) Live-cell time-lapse imaging of mid1-mNG or mid1–9A-mNG with sad1-mCherry. Images were acquired every 2 minutes and every other time point is shown. Scale bars = 2 μm. (A) Time zero is the first frame of SPB separation or for (B) the first frame that septation was detected in the DIC image. (C) Live-cell imaging of mid1–9A-mNG blt1-mCherry. Scale bar = 2 μm. (D) Live-cell SIM imaging of mCherry-cdc15 mid1–9A-mNG. Cells were loaded into a device to vertically position the cells [48]. Scale bar = 2 μm. (E) Live-cell imaging of mid1-mNG and mid1–9A-mNG in wildtype or imp1Δ. Scale bar = 5 μm. (F, top) Five representative line scans from mid1-mNG and mid1-mNG imp1Δ. (F, bottom) Five representative line scans from mid1-mNG imp1Δ and mid1–9A-mNG. Lines were drawn through the nucleus along the short axis of the cell. The approximate position of the cell cortex is indicated. Also see Figure S3 and Table S1.

Mid1 shuttles in and out of the nucleus via its multiple nuclear localization and export sequences [7,11]. To address the possibility that Mid1–9A has increased cortical localization because it cannot be imported into the nucleus, we constructed a strain defective in Mid1 nuclear import. Purification of Mid1-TAP followed by mass spectrometric analysis identified Imp1 (protein coverage, 39.1%), one of two S. pombe importin-α proteins [28], as a co-purifying protein suggesting that Imp1 might mediate Mid1 nuclear import. Indeed, deletion of imp1 inhibited Mid1’s nuclear accumulation (Figure 2E). Line scans through the middle of the short axis revealed a lack of medial nuclear signal for mid1-mNG imp1Δ cells (Figure 2F, top). In imp1Δ cells, Mid1-mNG did not however display increased cortical localization similar to Mid1–9A-mNG (Figure 2E-F, bottom). Furthermore, deleting imp1 did not increase Mid1–9A-mNG cortical localization (Figure 2E). Therefore, Imp1 mediates Mid1 nuclear import and defects in Mid1 nuclear import do not significantly alter Mid1 cortical localization.

Mid1 phosphorylation influences cytokinesis dynamics and division site placement

To determine if the failure to dissociate Mid1 from the membrane affects cell division, we imaged mid1⁺ and mid1–9A using Rlc1-mNG and Sid4-mNG as markers of cytokinetic and mitotic events, respectively. Medial Rlc1 recruitment occurred earlier in mid1–9A cells, averaging ~17 minutes prior to SPB separation compared to ~6 minutes for wildtype (Figure 3A-B). mCherry-Cdc15, GFP-Myo2, and GFP-Rng2 were also recruited early (Figure S2A-B). In contrast, we did not detect precocious medial Cdc12-mNG in mid1–9A cells (Figure S2A-B) with the caveat that Cdc12 is a low abundance protein [29], which may fall below our detection limit. We also did not observe medial Plo1-mNG in mid1–9A (Figure S3C) indicating that unlike in wildtype [9], Plo1 may not be required for Mid1–9A to recruit early CR proteins. Despite early arrival of some CR proteins, the time from SPB separation to CR assembly (CR formation) was not different between strains (Figure 3B). However, the time between CR formation and onset of constriction (CR maturation) and constriction were significantly longer in mid1–9A (Figure 3B). Therefore, Sid2 phosphorylation of Mid1 contributes to timely completion of cytokinesis.

Figure 3. Sid2 phosphorylation of Mid1 modulates CR dynamics.

(A) Live-cell time-lapse imaging of mid1⁺ or mid1–9A expressing rlc1-mNG sid4-mNG. Images were acquired every 2 minutes and every other time point is shown. Time zero is the first frame with SPB separation. Scale bar = 2 μm. (B) Quantification of cytokinesis event timing for each strain. Measurements are from at least three biological replicates. WT CR formation vs. mid1–9A, p = 0.27. **p ≤ 0.01, ****p ≤ 0.0001, Student’s t-test. Error bars represent SEM. (C) The indicated strains were grown to log phase at 32°C and then fixed and stained with DAPI and Methyl Blue. The blue arrow indicates a cell with an off-centered septum. Scale bar = 5 μm. (D) Quantification of the ratio of the short to long daughter cell for each indicated strain. Black bars represent mean. n ≥ 100. ****p ≤ 0.0001, one-way ANOVA. Also see Figure S2 and Table S1.

Cdr2 is also a Sid2 substrate [18] and dispersed from nodes by SIN hyperactivation [4]. A mutant lacking two Sid2 phosphosites (Cdr2–2A) accumulates in nodes near the new cell end following cytokinesis instead of dispersing into the cytoplasm [18]. To determine whether combining Cdr2 and Mid1 Sid2 phosphosite mutants would cause division site mis-positioning, we measured septa position in cdr2–2A, mid1–9A and cdr2–2A mid1–9A cells. Neither single mutant had off-center septa but the double mutant did (Figure 3C-D). These data indicate that Cdr2 and Mid1 are both important substrates for Sid2 in resetting node proteins properly for the next round of cell division.

Sid2 phosphorylation of Mid1 regulates its association with Cdr2 and G2/M timing

We next examined intrinsic features and interaction partners of Mid1 to determine how Sid2 modulates Mid1 localization. Mid1 membrane localization, but not CR localization, requires a proposed amphipathic helix within its C2 domain (Figure 1A), and mutation of charged residues within the helix abolishes Mid1’s node localization (Figure S3A) [11]. When these mutations were introduced into Mid1–9A-mNG, it no longer localized to nodes (Figure S3A) or to the membranes lining the septum during CR constriction (Figure S3B). Dimerization of Mid1’s C2 domain increases Mid1’s avidity for membranes and dimerization can be disrupted with three point mutations (Mid1–3A) [30]. Mid1–3A-mNG had on average fewer nodes/cell at all cell lengths compared to wildtype and when the 9A mutant was combined with 3A (Mid1–9A-3A-mNG), the number of nodes/cell more closely resembled wildtype (Figure S3C-D). Thus, the Mid1–9A mutant also requires C2-mediated dimerization for robust node localization.

Two node components, Gef2 and Cdr2, each bind Mid1 at regions adjacent in primary sequence to Sid2 phosphorylation sites [9,10,20,31] (Figures 1A, S4A). To test whether Cdr2, Gef2, or a third node component, Blt1, are required for Mid1–9A-mNG node association, we analyzed Mid1–9A-mNG localization in gef2Δ, cdr2Δ and blt1Δ. Mid1–9A-mNG node localization was not altered in gef2Δ or blt1Δ but it was significantly reduced in cdr2Δ (Figure S4A). To determine if the increased amount of Mid1–9A-mNG in nodes reciprocally affected Cdr2 node levels, we analyzed Cdr2-mNG localization in mid1⁺ and mid1–9A. In wildtype cells, Cdr2-mNG localized to nodes during interphase and was diffuse through the cytoplasm of dividing cells, only sometimes faintly co-localizing with Rlc1-mCherry in ~18% of CRs (Figure 4A-B) [8,32]. Cdr2 node intensity was unaffected by mid1–9A (Figure S4B), consistent with Cdr2 acting upstream of Mid1 [20]. However, Cdr2-mNG colocalized with Rlc1-mCherry in ~75% of mid1–9A CRs (Figure 4A-B), suggesting that Mid1–9A may ‘drag’ Cdr2 into the CR. Thus, we tested whether Sid2-mediated Mid1 phosphorylation modulates Mid1-Cdr2 association. Cdr2 co-immunoprecipitated ~2x more Mid1–9A-myc₁₃ than Mid1-myc₁₃ (Figure 4C-D). Cdr2 also co-immunoprecipitated ~2x more Mid1-myc₁₃ from SIN-inactive sid2–250 arrested cells compared to asynchronous cells (Figure S4C-D). Thus, Sid2 phosphorylation antagonizes Mid1-Cdr2 association in nodes and acts to remove both proteins from the membrane.

Figure 4. Sid2 phosphorylation of Mid1 controls Mid1’s association with Cdr2.

(A) Live-cell images of cdr2-mNG rlc1-mCherry in mid1⁺ or mid1–9A cells. Arrow in bottom panel indicates Cdr2 colocalized with Rlc1. (B) Quantification of the percent of cells with Cdr2-mNG colocalized with Rlc1-mCherry in non-constricted CRs in mid1⁺ and mid1–9A cells. (C) Co-immunoprecipitation of Cdr2-flag and Mid1-myc₁₃ from a cdc25–22 G2 block. Anti-flag immunoprecipitates were resolved by SDS-PAGE and blotted for anti-flag and anti-myc₁₃. The asterisk denotes a non-specific band present in all samples. CDK was used as a loading control. A representative experiment is shown and (D) represents the quantification of two biological replicates. Mid1 levels were normalized to both Cdr2-flag and CDK levels. (E) Measurements of cell length at septation of wildtype (wt) and mid1–9A cells grown at the indicated temperatures. Magenta bars represent the mean. n ≥ 300 and from three biological replicates. ****p ≤ 0.0001, Student’s t-test. (F) Live-cell imaging of cdr2Δ mid1–9A cells. (G) Quantification of the septa phenotypes of the indicated strains from three biological replicates. n ≥ 150. (H) Measurements of the ratio of short to long daughter cells of the indicated strains. Magenta bars represent mean. ****p ≤ 0.0001, one-way ANOVA. (I) Live-cell time-lapse imaging of mid1-mNG rlc1-mCherry cdr2Δ strain at 25°C. Images were acquired every 4 minutes and every 8 minutes are shown. Time zero denotes the beginning of the movie. The blue line indicates the center of the cell in the mCherry channel. Scale bars = 5 μm. Also see Figure S4 and Table S1.

When imaging mid1–9A cells, we noticed that they reproducibly divided at a shorter cell length than wildtype, indicating that the G2/M transition was advanced (Figure 4E and S4E) [33]. Cdr2 not only recruits Mid1 to Type1 nodes but activates the G2/M transition via inhibiting Wee1 kinase activity [3] at nodes [34]. Because Cdr2 binds more Mid1–9A than Mid1, we tested whether Mid1–9A requires Cdr2 to advance the G2/M transition. We found that it did (Figure S4F-G). Cdr1 kinase, which also inhibits Wee1 [3] and localizes to Cdr2 nodes [34], was also required for Mid1–9A to influence mitotic onset (Figure S4F-G). The DYRK kinase Pom1 inhibits the G2/M transition and pom1Δ cells are also short [2]. Combining pom1Δ with mid1–9A cells did not decrease cell length further, indicating that Pom1 and Mid1 function in the same pathway to influence G2/M transition timing (Figure S4H). It will be interesting to better understand how enhanced Cdr2-Mid1 interaction influences this key cell cycle transition.

Sid2 phosphorylation of Mid1 is critical for proper division site placement in the absence of Cdr2

Cdr2 is critical for the establishment of Type1 interphase nodes and recruiting Mid1 for medial CR assembly [8,9,20]. To determine the consequence of allowing Mid1 to remain cortical in the absence of Cdr2-organized nodes, we examined cytokinesis in mid1–9A cdr2Δ cells. These cells exhibited an increased percentage of abnormal septa compared to either single mutant and wildtype cells (Figure 4F-G). They also had dramatically off-centered septa (Figure 4H). The mid1–9A cdr2Δ phenotype is distinct from that of mid1Δ (Figure 4G), and also distinct from mid1Δ cdr2Δ cells, which have a very high percentage of abnormal septa of different variety (Figure 4G and S4I). To determine how septa become off-centered in mid1–9A cdr2Δ, we performed live-cell time-lapse imaging (Figure 4I). We observed that these cells formed CRs off-center and/or tilted from a broad distribution of cytokinetic nodes rather than from medially organized nodes (Figure 4I). Thus, in the absence of nodes, SIN-mediated Mid1 phosphorylation is required to ensure correct division site positioning.

Conclusions

Mid1 joins a growing list of bonafide Sid2 substrates [18,35–40]. We discovered that 1) Sid2 phosphorylation of Mid1 modulates its membrane localization and that this is separable from the removal of Cdr2 from membranes by SIN signaling [4,18], 2) SIN-mediated removal of Mid1 from the division site is required for normal timing of cell division, 3) Mid1’s association with Cdr2 is inhibited by the SIN and influences G2/M transition timing, and 4) SIN phosphorylation of Mid1 is key to proper division plane positioning, particularly in the absence or mis-regulation of node formation.

When cytokinetic nodes form, Mid1 transfers from Type1 to Type2 nodes, presumably by dissociating with Cdr2 and establishing new connections [8]. Our results suggest that Mid1–9A does not dissociate from Cdr2 at this stage. As a result, Mid1–9A remains associated with the division site membrane and available to participate prematurely in establishing the future division site. Thus, we posit that a key restructuring step in the CR does not occur if Mid1 is not phosphorylated by Sid2. Our findings on Mid1 spatial re-organization at this stage are consistent with the reported difference in organization between Mid1 and other cytokinesis node components in assembled CRs using fPALM [41].

Surprisingly, cdr2 deletion, which abrogates node formation, led to more severe defects in division site placement when combined with mid1–9A than did combining Mid1–9A with a Cdr2 mutant that persists on the cortex. However, unlike what happens in the absence of Cdr2, the Cdr2–2A mutant can organize Mid1 into medial nodes [18]. Together, these findings indicate that both node assembly during interphase and node disassembly during mitosis are important for proper division site placement in S. pombe.

Anillin proteins play a conserved role in promoting cytokinesis in eukaryotes [1] and mechanisms of anillin and Mid1 accumulation at the membrane and ring both depend on phosphorylation [9,42]. Now we show that Mid1 removal from the membrane is also phosphorylation-dependent. Because altering the levels of anillin at the membrane can speed up or slow down steps of cytokinesis [43–46], timely turn-over of anillin proteins at the membrane via phosphoregulation may be a general requirement for proper division site placement in eukaryotes.

It is noteworthy that several Sid2 substrates (Cdr2, Klp2, and now Mid1) are involved in re-establishing an interphase state following cell division [18,39]. Such coordinated reorganization of cytoskeletal-related processes is a general theme in Hippo pathway signaling as well [47]. Therefore, our results may also provide insights to how this conserved pathway coordinates processes at the end of cell division in higher eukaryotes.

STAR METHODS

Contact for Reagent and Resource Sharing

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Kathleen L. Gould (Kathy.gould@vanderbilt.edu)

Experimental Model and Subject Details

Fission Yeast

S. pombe strains used are listed in the Table S1. Strains were prepared by using standard fission yeast genetic techniques. Cells were cultured in rich medium YES (5 g/L yeast extract, 30 g/L glucose, 225 mg/L adenine) with supplements until mid-log phase at 25°C or 32°C for analysis. Mid1 and Mid1–9A were each C-terminally tagged with mNeonGreen (mNG) and produced from the endogenous mid1⁺ locus.

Method Details

Protein Expression and Purification

Mid1-TAP was purified from nda3-KM311 arrested cells [49]. To identify phosphorylation sites, TAP samples were analyzed by MudPIT after in solution digestion with trypsin and then LC-MS/MS mass spectrometric analysis. The obtained mass spectra were filtered by ScanDenser (Vanderbilt University Mass Spectrometry Research Center) and searched against the Sanger Institute S. pombe database using SEQUEST (Thermo Finnigan).

Mid1(1–578) and His₆-Mid1(1–578-S110A, S127A, S218A, S227A, S329A, S432A, S452A, S464A, S531A) (9A) were cloned into pET15b for expression as an His₆ fusion. Proteins were induced in Escherichia coli Rosetta2(DE3)pLysS cells with 0.4 mM IPTG overnight at 18°C. Proteins were purified on c Omplete His-Tag resin (Roche) according to the manufacturer’s protocols.

Protein Methods

Cell pellets were snap-frozen in dry ice–ethanol baths. Lysates were prepared using a Fastprep cell homogenizer (MP Biomedicals). Immunoprecipitations were performed in NP-40 buffer (6 mM Na₂HPO₄, 4 mM NaH₂PO₄, 1% NONIDET P-40, 150 mM NaCl, 2 mM EDTA, 50 mM NaF, 4 ug/l leupeptin, 0.1 mM Na₃VO₄, 1 mM PMSF, 2mg/ml aprotinin, 2 mM Benzamidine). For Figure 4B and S4C, four 30 OD pellets were used per sample. Protein samples were resolved by SDS–PAGE and transferred to polyvinylidene fluoride (PVDF) membrane (Immobilon-FL; Fisher Scientific). Anti-myc (9E10; Thermo Fisher Scientific), anti-FLAG (M2; Sigma) and anti-Cdc2 (PSTAIRE; Sigma) were used in immunoprecipitations and/or as a primary antibody in immuno-blotting. Secondary antibodies were conjugated to IRDye800 (LI-COR Biosciences). Blotted proteins were detected via an Odyssey Classic (LI-COR Biosciences). For gel shifts, immunoprecipitates from denatured lysates were treated with λ-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 befo re the reaction was quenched with 5 μl of 5x SDS sample buffer.

Kinase reactions were performed in NDR protein kinase buffer (25 mM MOPS, pH 7.2, 60 mM β-glycerol phosphate, 15 mM MgCl₂, 1 mM DTT, 0.1 mM sodium vanadate, 1% NP-40) with 10 M cold ATP, 3 μCi of [32P]ATP, and immunoprecipitated Sid2-myc kinase in 20 μl reactions that were incubated at 30°C for 30 min with shaking. Reactions were quenched by the addition of 5 μl of 5x SDS sample buffer. Proteins were separated by SDS–PAGE and transferred to PVDF membrane, and phosphorylated proteins were visualized by autoradiography.

Microscopy Methods

Live-cell images of S. pombe cells were acquired using a Personal DeltaVision (Applied Precision) that includes a microscope (IX71; Olympus), 60× NA 1.42 Plan Apochromat and 100× NA 1.40 U Plan S Apochromat objectives, fixed and live-cell filter wheels, a camera (CoolSNAP HQ2; Photometrics), and softWoRx imaging software (Applied Precision). z sections were spaced at 0.5 μm. Images were acquired at indicated temperature and cells were imaged in YES media. Images were deconvolved with 10 iterations.

Intensity measurements were made with ImageJ software [50]. For all intensity measurements, the background was subtracted by creating a region of interest (ROI) in the same image where there were no cells. The raw intensity of the background 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. All images used for quantification were not deconvolved. For Mid1 node number and intensity measurements, the first three z-slices were either max or sum projected, respectively. Mid1 node number/cell was quantified by applying the same scaled Fire look-up table to all images and only counting nodes that had a yellow, orange or white intensity. Nodes with only purple or red intensity were not counted. For Mid1 intensity measurements, an ROI in the middle of the cell was measured as described above. For measuring the nuclear to cytoplasmic ratio of Mid1 signal, whole cell fluorescence intensity and nuclear intensity were both measured from a single medial z-slice. Cytoplasmic signal was determined by subtracting nuclear intensity from whole cell fluorescence intensity.

For nuclei and cell wall imaging, cells were fixed in 70% ethanol for at least 30 min before DAPI and Methyl Blue staining. To quantify off-center septa, the coordinates of the cell tips and septum were logged. Lengths of the shorter and longer cell were calculated from these coordinates and reported as a ratio.

Structured illumination microscopy was performed on a Nikon SIM microscope equipped with an EMCCD camera, 488 nm and 561 nm lasers, 60× 1.27 NA objective and NIS Elements imaging software (Nikon).

Quantification and Statistical Analysis

Calculations of mean, standard error of the mean (SEM), and statistical significances were performed with Prism 6.0 (GraphPad Software). Significance was defined by a p value equal to or less than 0.05. For all data following a normal distribution, a Student’s two-sample unpaired t test (Figure 1I, 3B and 4E) or an ANOVA test was used with Tukey’s post-hoc analysis (Figure 3D). For samples with a minor deviation from a normal distribution (Figures 4H and S4F), an ANOVA was also used with Tukey’s post-hoc analysis. Sample sizes and the numbers of replications are included in the figure graphs or legends. Where indicated, n represents the number of cells used for quantification. For all image analyses, no raw data were excluded with the exception of cells that were not in focus or if a cell moved during imaging.

Highlights:

• Anillin-like Mid1 is a substrate of the SIN kinase Sid2

• SIN signaling controls Mid1’s membrane localization

• Sid2 temporally regulates Cdr2-Mid1 association

• Cdr2 and Mid1 are key SIN substrates for resetting cell division site placement