Phosphoinositide-mediated ring anchoring resists perpendicular forces to promote medial cytokinesis

To divide, many eukaryotes assemble and constrict an actin- and myosin-based contractile ring (CR; Cheffings et al., 2016) that is anchored to the plasma membrane (PM; Gould, 2016). Despite decades of work on how the division plane is selected and the CR assembles (Pollard and Wu, 2010; Bohnert and Gould, 2011; Goyal et al., 2011; Lee et al., 2012; Rincon and Paoletti, 2016), mechanisms of CR–PM anchoring remain incompletely understood.

One factor implicated in CR–PM attachment is PM lipid composition. Phosphatidylinositol (PI)-4,5-bisphosphate (PI(4,5)P₂) and PI5 kinases are enriched at the cleavage furrow of mammalian cells, and when PI(4,5)P₂ is depleted, cortical actin separates from the PM (Emoto et al., 2005; Field et al., 2005). CR components such as RhoA, anillin, and MgcRacGAP bind PI(4,5)P₂ (Cauvin and Echard, 2015); therefore, they may mislocalize when PI(4,5)P₂ levels are reduced. A decrease in cortically bound anillin may be particularly detrimental to CR–PM attachment because anillin binds components of the CR and PI(4,5)P₂ (Piekny and Glotzer, 2008; D’Avino, 2009; Sun et al., 2015). In human and Drosophila melanogaster cells, anillin loss causes cleavage furrow oscillations (Kechad et al., 2012). However, the exact combination of molecules involved in CR detachment when PI(4,5)P₂ is depleted remains to be determined.

Schizosaccharomyces pombe build a medial CR early in mitosis (Marks and Hyams, 1985; Kitayama et al., 1997). How the CR remains anchored until constriction is not yet clear, although several players have been implicated. One is the Fes/CIP4 homology Bin-Amphiphysin-RVS (F-BAR) protein Cdc15: when cdc15 expression is repressed or Cdc15 oligomerization is disrupted, the CR can slide along the PM and disassemble (Arasada and Pollard, 2014; McDonald et al., 2015). Paxillin Pxl1 also plays a role in CR anchoring and integrity, indicated by CR sliding and splitting during anaphase in pxl1Δ (Ge and Balasubramanian, 2008; Cortés et al., 2015). Another factor is the cell wall: loss of β(1,3)glucan (Muñoz et al., 2013) or loss of the integral membrane protein Sbg1 (Davidson et al., 2016; Sethi et al., 2016) result in CR sliding and instability, suggesting that cell wall–PM linkage is important for CR maintenance. Finally, the microtubule post-anaphase array ensures a medial CR during a cytokinesis arrest (Pardo and Nurse, 2003). In each of these situations, CR sliding is observed in only a fraction of cells (Pardo and Nurse, 2003; Arasada and Pollard, 2014; Cortés et al., 2015; McDonald et al., 2015), indicating that multiple mechanisms contribute to CR anchoring. Consistent with this, combined repression of pxl1 with a hypomorphic cdc15 allele results in exacerbated CR sliding (Cortés et al., 2015).

In this study, we define a distinct mechanism that anchors CRs during anaphase, explaining why cells lacking S. pombe efr3 divide asymmetrically (Chen et al., 2015). In Saccharomyces cerevisiae, Efr3 and its partner Ypp1 form a platform at the PM for Stt4, a PI4 kinase, which regulates phosphoinositide (PIP) composition and supports endocytosis (Baird et al., 2008). Similarly, human homologues of Efr3 and Ypp1 (EFR3A/B and TTC7) scaffold a PI4 kinase type-IIIα at the PM (Nakatsu et al., 2012).

We find that S. pombe lacking properly positioned Stt4 have altered PM PIPs. These cells form CRs in the cell middle that can then slide toward one end in a directed manner. CR sliding in efr3Δ requires the type V myosin Myo51, indicating for the first time that the CR is subject to perpendicular forces in addition to being under constrictive tension (Proctor et al., 2012) and that these forces can dislodge the CR from the cell center. Thus, PM PIP composition contributes to CR anchoring, promoting proper septum positioning and ensuring accurate genome segregation.

We previously observed that a high percentage of efr3Δ divide asymmetrically (Fig. 1 A), sometimes resulting in the cutting of chromosomes by off-center septa (Chen et al., 2015). To determine whether this is caused by altered PM PIP composition, we first determined whether Efr3 colocalizes with Stt4 and Ypp1 in S. pombe by analyzing the localization of three distinct pairs of these proteins tagged with mCherry and mNeonGreen (mNG) or GFP. Each pair colocalized on the PM in a punctate pattern (Fig. 1 B), resembling the PI kinase patch localization of the S. cerevisiae Stt4 complex (Baird et al., 2008). The PM enrichment of S. pombe Stt4 and Ypp1, but not their levels, depended on Efr3 (Fig. 1 C). Efr3 coimmunoprecipitated with Ypp1 (Fig. S1 A), and both Ypp1 and Stt4 were identified in an Efr3-TAP by liquid chromatography–tandem mass spectrometry analysis (Fig. S1 B and Table S2), indicating that these proteins associate in S. pombe. To further study the influence of Stt4 on septa positioning, we attempted to construct stt4Δ and ypp1Δ but found that these genes are essential (Fig. S1, C and D). However, endogenously tagged GFP-stt4 displayed off-center septa, indicating that although GFP-Stt4 localizes correctly to the PM, it is likely to be a hypomorphic allele (Fig. 1, A and B). These data establish that proteins of the Stt4 complex are important for medial division.

Stt4 phosphorylates PM PI to produce PI-4-phosphate (PI4P), which can be further modified to PI(4,5)P₂. Therefore, disruption of PI4 kinases results in a reduction of both PI4P and PI(4,5)P₂ (Audhya and Emr, 2002; Baird et al., 2008; Nakatsu et al., 2012). The PI(4,5)P₂ sensor GFP-2×PH(PLCδ) (Stefan et al., 2002) was reduced at the cell cortex and the division site in efr3Δ compared with WT (Fig. 1 D), indicating that PIP PM abundance is reduced in efr3Δ. In accordance with this, overexpression of GFP-2×PH(PLCδ), expected to sequester PI(4,5)P₂, resulted in off-center septa (Fig. 1 E).

We next addressed how off-center septa arise in efr3Δ. Because septum position is dictated by CR position (Marks and Hyams, 1985; Marks et al., 1986), we reasoned that either the CR forms off-center or it slides from its original medial position. To distinguish between these possibilities, we imaged WT and efr3Δ expressing CR (Rlc1-GFP) and spindle pole body (Sid4-GFP) markers. In WT, the CR formed in the cell center and maintained this position during cytokinesis (Fig. 2 A). In efr3Δ, the CR formed in the cell center but slid from its original position while remaining perpendicular to the cell’s long axis (Fig. 2, A–C). Temporal progression through cytokinesis was unchanged in efr3Δ compared with WT (Fig. S2, A and B). CR sliding occurred during anaphase B, after the CR formed in early mitosis, but stopped before or coincidently with CR constriction (Fig. 2, A and D). This indicates that the CR cannot slide in efr3Δ once septation begins, likely because septum formation locks the CR in position (Muñoz et al., 2013; for review, see Willet et al., 2015).

CR sliding, indicative of a CR anchoring defect, was observed when oligomerization of Cdc15’s F-BAR domain was prevented or when cdc15 expression was repressed (Arasada and Pollard, 2014; McDonald et al., 2015). To determine whether Cdc15-mediated CR anchoring involves Efr3, we compared CR sliding events in efr3Δ and cdc15–core 4A, a mutant that specifically impairs membrane binding (McDonald et al., 2015). Alone, cdc15–core 4A displayed CR sliding. When combined with efr3Δ, the frequency of CR sliding events was increased compared with each single mutant (Fig. 2, A and B). Also, CRs slid farther in the double mutant, indicated by the lower mean ratio of short to long cells at septation (Fig. 2 C), ultimately leading to growth defects (Fig. 2 E). Therefore, Cdc15- and Efr3-mediated CR anchoring are independent mechanisms that maintain central CR positioning. Mutants of pxl1 also display CR sliding (Cortés et al., 2015); however, pxl1Δ efr3Δ was inviable (Fig. S2 D), suggesting that Pxl1 contributes to CR anchoring independently of Efr3. Because efr3Δ does not change the kinetics of cytokinesis (Fig. S2, A and B) as defects in the β-glucan enzymes Bgs1 and Bgs4 do (Muñoz et al., 2013; Davidson et al., 2016; Sethi et al., 2016), Efr3-dependent anchoring appears to be an independent mechanism from cell wall anchoring as well.

Because efr3Δ has reduced levels of PM PI(4,5)P₂, we hypothesized that the cortical enrichment of proteins with membrane-binding domains (F-BAR, Pleckstrin homology [PH], Phox homology domain [PX], or C2) would be diminished in efr3Δ compared with WT. Consistent with Cdc15 acting independently of Efr3 and interacting with a wide range of anionic phospholipids (McDonald et al., 2015), we found no difference in Cdc15 CR intensity (Fig. 3, A and B). The localizations of many other membrane-binding proteins, such as Pob1 (Fig. S2 E), were also unaltered in efr3Δ (Fig. 3 A), suggesting that they do not rely upon PI4P or PI(4,5)P₂. However, we identified three PH domain–containing proteins with reduced PM localization in efr3Δ compared with WT. The RhoGEF Rgf1 and Cdc42 GEF Scd1 were reduced at the division site without any reduction in total protein levels (Fig. 3, C and D). Opy1, encoded by the ORF SPCPB16A4.02c, is normally enriched at the PM but was diffusely localized in efr3Δ (Fig. 3 E). Opy1 contains two PH domains, and the S. cerevisiae orthologue Opy1 is implicated in sensing PI4P at the PM and inhibiting the PI5 kinase Mss4 (Its3 in S. pombe; Ling et al., 2012). Thus, it may be a collective reduction of several proteins at the cortex that compromises CR–PM attachment in efr3Δ.

Off-center septa in S. pombe are observed when CRs slide and also when they assemble off center. Mid1 and Pom1 dictate CR positioning (Rincon and Paoletti, 2016); cells lacking either divide asymmetrically because of misplaced CRs (Chang et al., 1996; Sohrmann et al., 1996; Bähler and Pringle, 1998; Bähler et al., 1998a). In mid1Δ, CRs assemble at random positions and angles along the cortex but eventually coalesce into an orthogonal CR (Huang et al., 2008); although if CRs form within the curved cell pole of mid1–18, they can slide toward the tip, decreasing in diameter (Mishra et al., 2012). To test whether the Mid1 and Pom1 cues are influenced by PM composition, given that both proteins bind membrane PIPs (Celton-Morizur et al., 2004; Hachet et al., 2011), we combined mid1Δ or pom1Δ with efr3Δ and analyzed CR dynamics. We scored whether CRs formed off-center and whether fully formed CRs slid during anaphase. In our experiments, mid1Δ rarely formed CRs at the extreme cell tip that slid. As previously observed, mid1Δ and pom1Δ form off-center CRs, but we did not detect a significant number of subsequent sliding events (Chang et al., 1996; Sohrmann et al., 1996; Bähler and Pringle, 1998). Consistent with our finding that initial CR placement is not altered in efr3Δ, the localizations of PM-binding proteins Mid1 and Pom1 were not influenced by deletion of efr3 (Fig. 3 A). In combination with efr3Δ, mid1Δ and pom1Δ CRs formed off-center and slid during anaphase, resulting in septa that were farther off-center than single mutants (Fig. 4, A–C). These combinations also led to significant growth defects (Fig. 4 D), most likely a result of chromosomes being cut by off-center septation.

We next considered whether CR sliding events in efr3Δ are enabled by diminishing cell circumference from the cell middle toward one end, similar to CRs in spheroplasts, which move along the cortex while constricting (Mishra et al., 2013). This seemed unlikely, however, because efr3Δ has normal morphology (Chen et al., 2015) and CRs slide only 1–2 µm (Fig. 2 A), not approaching the region of curvature at the hemispherical cell ends (Atilgan et al., 2015). Also, pos5Δ, which is tapered at one cell end (Hayles et al., 2013), does not have off-center septa, indicating that a CR does not automatically slide toward a tapered end (Fig. S2 C). The CR width, as a proxy of cell diameter, also does not decrease during a sliding event in efr3Δ, which would be expected if the circumference of the cell changed (Δ width = 0.045 µm ± 0.039 µm [SEM; n = 19]). Furthermore, in pos5Δ efr3Δ, there was no worsening of the off-center septa phenotype compared with efr3Δ (Fig. S2 C), and sliding CRs had no bias toward the tapered end of the cell (54% to the tapered end vs. 46% to the nontapered end), similar to efr3Δ, where the CR is equally likely to slide to the old or new end of the cell (48% to the old end vs. 52% to the new end). We conclude that CR sliding events in efr3Δ are not dictated by cell geometry.

Because CR sliding in spheroplasts and cdc15-defective cells depends on the type II myosins Myo2 and Myp2, respectively (Mishra et al., 2013; Arasada and Pollard, 2014), we tested whether myosin-generated force is required for CR sliding in efr3Δ. Neither Myp2, the type I myosin Myo1, nor the type V myosin Myo52 were necessary for CR sliding in efr3Δ (Fig. 5 A). To test whether CR sliding depends on Myo2, we attempted to combine efr3Δ with the temperature-sensitive myo2-E1 allele (Balasubramanian et al., 1998), but these alleles were synthetically lethal (Fig. S3 A). Instead, we used a myo2-E1 GFP-stt4 combination. GFP-stt4 (GFP-stt4 is hypomorphic) and myo2-E1 each had off-center septa at 32°C, and the combination resulted in more cells with off-center septa that were even farther away from center (Fig. S3 B). Live-cell imaging revealed that GFP-stt4 CRs slid during anaphase, whereas myo2-E1 CRs formed off-center but did not slide (Fig. S3 C). GFP-stt4 myo2-E1 formed CRs off-center that then also slid along the cortex (Fig. S3 C). Thus, CR sliding in efr3Δ does not depend on Myo2.

In contrast, deletion of the type V myosin Myo51 eliminated the efr3Δ off-center septa phenotype (Fig. 5 A). CR sliding events no longer occurred in myo51Δ efr3Δ (Fig. 5, B and C), and the mean ratio of short to long cells was significantly higher than in efr3Δ (Fig. 5 A). As expected given that Stt4 binds Efr3, the GFP-stt4 off-center septa phenotype is also Myo51 dependent (Fig. S3 B). None of the myosins were necessary for CR sliding in cdc15–core 4A or pxl1Δ mutants (Fig. S3, D and E), consistent with the genetic evidence that Cdc15-, Pxl1-, and Efr3-dependent CR sliding events occur through independent mechanisms (Figs. 2 E and S2 D; Cortés et al., 2015). Indeed, CR sliding in cdc15 and pxl1Δ mutants as well as in β-glucan synthase mutations may occur because of structural instability of the CR rather than a directed movement of the CR along the cortex (Balasubramanian et al., 1998; Arai and Mabuchi, 2002; Wachtler et al., 2006; Ge and Balasubramanian, 2008; Hachet and Simanis, 2008; Roberts-Galbraith et al., 2009; Laporte et al., 2011; Muñoz et al., 2013; Arasada and Pollard, 2014; McDonald et al., 2015; Davidson et al., 2016; Sethi et al., 2016).

Myo51 contains an N-terminal motor head domain and a C-terminal tail domain necessary for CR localization (Wang et al., 2014). By testing Myo51 N- and C-terminal truncations in efr3Δ, we found that both the N-terminal head and C-terminal tail of Myo51 are necessary for CR sliding (Fig. S3 F), suggesting that Myo51 tail binding to CR components and the ability to walk along actin filaments (Wang et al., 2014) are both required to move the CR.

The necessity of Myo51 force generation for CR movements in efr3Δ provides strong evidence for the existence of forces on the CR perpendicular to the cell axis. Such forces could be involved in the splitting of CRs observed in some mutants, e.g., pxl1Δ (Ge and Balasubramanian, 2008) and sbg1–3 (Sethi et al., 2016). We hypothesize that perpendicular forces are balanced in a WT cell and/or that CR–PM attachments are sufficient to resist these forces so that the CR remains in its central position. In efr3Δ, force imbalances may arise that cannot be stabilized, resulting in Myo51-dependent CR sliding along the cell axis. In support of this hypothesis, mean squared displacement (MSD) analysis of sliding efr3Δ CRs shows a statistically significant drift velocity term, suggesting directed transport of the CR (Fig. S3 G; v = 1.35 nm/s; 95% confidence interval, 1.32–1.37 nm/s). No such transport term was measured for WT rings, which exhibit minimal changes in position over the imaging period (Fig. S3 G; also see the MSD analysis section of Materials and methods). A possible explanation for this behavior is that Myo51 in the CR associates with longitudinal actin cables as well as F-actin within the CR, pulling the CR along cables when PM anchoring is weakened. In support of this, Myo51 has been shown to play a role in the medial accumulation of actin cables during cytokinesis (Huang et al., 2012), and actin cables can be seen in proximity to sliding CRs in efr3Δ (Fig. 5 D).

Altogether, our data reveal a novel CR anchoring mechanism that depends on a conserved PM-localized PI4 kinase complex (Fig. 5 E). An ensemble of proteins sensitive to correct PM PIP composition synergize with Cdc15-, Pxl1-, and cell wall–dependent anchoring to promote stable CR placement and faithful segregation of the genetic material during cell division. A cytokinetic function for Stt4 or its human orthologues has not been previously reported, but a related PI4 kinase is important for Drosophila cytokinesis (Brill et al., 2000; Eggert et al., 2004). Given the large number of lipid-binding proteins at the cell division site of many eukaryotes, PI4 kinases are likely to be a generally important factor for CR–PM adhesion.

S. pombe strains (Table S1) were grown in yeast extract (YE). efr3⁺, ypp1⁺, opy1⁺, scd1⁺, and rgf1⁺ were tagged at the 3′ end of their ORFs with TAP:kan^R, Flag₃:kan^R, HA₃:hyg^R, mCherry:kan^R, mNG:kan^R, or mNG:hyg^R using pFA6 cassettes as previously described (Wach et al., 1994; Bähler et al., 1998b). 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. Fusion proteins were expressed from their native promoters at their normal chromosomal loci unless otherwise indicated. Strains producing myo51 truncations were obtained from Dr. J.-Q. Wu (Wang et al., 2014).

For growth assays, cells were grown to log phase at 25°C in YE, 10 million cells were resuspended in 1 ml of water, and 10-fold serial dilutions were made. 2.5 µl of each dilution was spotted on YE plates, and the plates were incubated at the indicated temperatures.

Stt4 was N-terminally tagged with GFP at the endogenous locus using a Cre–loxP method as described previously (Werler et al., 2003). A cassette with the sequence that encodes GFP, the sup3–5 selection marker, and a temporary promoter (nmt1) were integrated at the 5′ end of the stt4 ORF. Next, the selection marker and temporary promoter were removed with Cre recombinase, resulting in the insertion of sequences encoding GFP at the 5′ end of stt4 under the normal promoter.

To express GFP-2×PH(PLCδ) in cells, the medium-strength cdc2 promoter (Meddins et al., 1993) was PCR amplified from S. pombe genomic DNA, and GFP-2×PH(PLCδ) was PCR amplified from plasmid pRS426 (Stefan et al., 2002). The two fragments were cloned into S. pombe integration vector pJK148 using Gibson assembly. This construct was linearized and inserted into the S. pombe leu1 locus by a lithium acetate method (Keeney and Boeke, 1994).

To overexpress GFP-2×PH(PLCδ), sequences encoding it were PCR amplified from plasmid pRS426 and cloned into pREP81 (Basi et al., 1993). This construct was introduced into cells by sorbitol transformation. Cells were fixed in 70% ethanol after induction of expression for 24 h at 32°C.

Live-cell images of S. pombe cells were acquired using a personal DeltaVision microscope system (Applied Precision Ltd.) that includes an IX71 microscope (Olympus), 60× 1.42 NA Plan Apochromat and 100× 1.40 NA U-Plan S-Apochromat objectives, a Coolsnap HQ2 camera (Photometrics), and SoftWoRx imaging software (Applied Precision Ltd.). Images were acquired at 25–29°C, and cells were imaged in YE media. Images in figures are either single slices or maximum-intensity projections of z sections spaced at 0.2–0.5 µm. Images used for quantification were not deconvolved. Other images not used for fluorescence quantification were deconvolved with 10 iterations. Time-lapse imaging was performed on cells in log phase on a YE agar pad at 32–36°C with the exception of LifeAct-mCherry cdc25-22 efr3Δ, where cells were shifted to 36°C for 4 h and then imaged at 25°C.

Intensity measurements were made with ImageJ software (National Institutes of Health). For all intensity measurements, the background was subtracted by creating a region of interest (ROI) in the same image where there were no cells (Waters, 2009). 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 (Waters, 2009). For CR intensity quantification, an ROI was drawn around the CR and measured for raw integrated density.

All cells were grown to log phase at 32°C before fixation. For nuclei and cell wall imaging, cells were fixed in 70% ethanol for 30 min before DAPI and methyl blue staining.

For quantification of ring sliding, a line was drawn through the fully formed CR marked by Rlc1-GFP using ImageJ software. Any movement of the CR away from the original line placement during the entire length of imaging was scored as a ring sliding event. For ring sliding distances, fixed cells stained for nuclei and cell walls were imaged. 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.

Time-lapse images of a medial z slice of a strain expressing a CR and spindle pole body marker were acquired every 10 s, registered for both differential interference contrast and fluorescent channels using the ImageJ plugin Image Stabilizer (Li, 2008), and then merged together to determine the position of the ring relative to the cell boundary. Cytokinetic ring positions were recorded for 17 WT and 18 efr3Δ mutant cells, beginning when the CR was fully formed. Individual trajectories were obtained by reslicing the time-lapse images to a 3-pixel-wide line along the cell boundary and tracking the position of the ring with subpixel resolution using the ImageJ plugin TrackMate (Tinevez et al., 2017). Total absolute displacements for WT trajectories over a 300-s period were measured to be 83 ± 66 nm (standard deviation; n = 34 tracks from 17 cells), which is within the spatial resolution of the imaging (pixel size, 106 nm). For efr3Δ mutant cells, one-dimensional position over time data were used to perform MSD analysis for the first 300 s of the data using MATLAB (R2016b; MathWorks) and the msdanalyzer tool as previously described (Tarantino et al., 2014). MSD data for efr3Δ mutant cells were fit by the second-degree polynomial function:

where $D$ is the diffusion coefficient, $v$ is the drift velocity, and $σ$ is the noise term. The fitting procedure yielded:

r² = 99.90%.

Cell pellets were snap frozen in dry ice–ethanol baths. Lysates were prepared using a Fastprep cell homogenizer (MP Biomedicals). Immunoprecipitations were performed as previously described (Gould et al., 1991) in NP-40 buffer for native lysates. Protein samples were resolved by SDS-PAGE and transferred to PVDF membrane (Immobilon P; EMD Millipore). Anti-HA (12CA5), anti-Flag (M2; Sigma-Aldrich), anti-GFP (Roche), anti-Cdc15 serum (Roberts-Galbraith et al., 2009), or anti-PSTAIRE (Cdc2; Sigma-Aldrich) were used in immunoprecipitations and/or as primary antibodies in immunoblotting. Secondary antibodies were conjugated to IRDye680LT or IRDye800 (LI-COR Biosciences). Blotted proteins were detected via Odyssey (LI-COR Biosciences).

Purification of Efr3-TAP and subsequent identification of interacting proteins by mass spectrometry were performed as previously described (Gould et al., 2004; Chen et al., 2013; Elmore et al., 2014) with the following changes: a newer version of Scaffold (Scaffold v4.4.1.1; Proteome Software) was used, and the minimum peptide identification probability was changed to 95.0%.

All ANOVA statistical analyses used Tukey’s post hoc analysis.

Fig. S1 shows that Efr3 and Ypp1 coimmunoprecipitated, that Ypp1 and Stt4 were purified from an efr3-TAP, and that stt4 and ypp1 are essential genes. Fig. S2 shows that efr3Δ does not have altered CR kinetics, that pos5Δ does not have off-center septa, and that efr3Δ pxl1Δ is synthetically lethal. Fig. S3 shows efr3Δ and myo2-E1 are synthetically lethal, that efr3Δ off-center septa do not depend on Myo2, that cdc15–core 4A and pxl1Δ CR sliding do not depend on a myosin, that the N and C termini of Myo51 are necessary for CR sliding in efr3Δ, and MSD analysis of efr3Δ sliding CRs. Table S1 includes S. pombe strains used in this study, and Table S2 contains the mass spectrometry results identified in an Efr3-TAP pulldown.

We thank Drs. J. Beckley, N. McDonald, and S. Cullati for critical reading of this manuscript and Dr. J.-Q. Wu for strains.

We are grateful for the following support: National Institutes of Health grants GM101035 (to K.L. Gould), GM119552 (to M. Zanic), and T32GM008554-21 (to C.E. Snider) as well as American Heart Association grant 14PRE19740000 (to A.H. Willet). M. Zanic acknowledges the support from the Human Frontier Science Program and the Searle Scholars Program.

The authors declare no competing financial interests.

Author contributions: C.E. Snider, A.H. Willet, M. Zanic, and K.L. Gould were responsible for the experimental design. C.E. Snider, A.H. Willet, J.-S. Chen, and G. Arpağ performed all of the experiments. C.E. Snider, A.H. Willet, and K.L. Gould wrote the manuscript with contributions from J.-S. Chen, G. Arpağ, and M. Zanic.