Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Multistep and multimode cortical anchoring of tea1p at cell tips in fission yeast

Multistep and multimode cortical anchoring of tea1p at cell tips in fission yeast The EMBO Journal (2005) 24, 3690–3699 & 2005 European Molecular Biology Organization All Rights Reserved 0261-4189/05 | | THE THE www.embojournal.org EMB EMB EMBO O O JO JOU URN R NAL AL Multistep and multimode cortical anchoring of tea1p at cell tips in fission yeast skeleton (Hayles and Nurse, 2001; Chang and Peter, 2003). Hilary A Snaith, Itaru Samejima After a variety of stresses, tea1D cells become bent or and Kenneth E Sawin* branched (Verde et al, 1995; Mata and Nurse, 1997; Niccoli Wellcome Trust Centre for Cell Biology, University of Edinburgh, et al, 2003; Sawin and Snaith, 2004), in part because micro- Edinburgh, UK tubules are unable to specify positional information to the cell cortex (Sawin and Nurse, 1998; Feierbach et al, 2004; The fission yeast cell-polarity regulator tea1p is targeted Sawin and Snaith, 2004). In addition, during steady-state to cell tips by association with growing microtubule ends. growth, tea1D mutants exhibit a mostly monopolar growth Tea1p is subsequently anchored at the cell cortex at pattern, unlike wild-type cells, which exhibit monopolar cell tips via an unknown mechanism that requires both growth early in the cell cycle and bipolar growth later in the tea1p carboxy-terminus and the membrane protein the cell cycle (Mitchison and Nurse, 1985; Mata and Nurse, mod5p. Here, we show that a tea1p-related protein, 1997; Glynn et al, 2001). tea1D mutants are also defective tea3p, binds independently to both mod5p and tea1p, in the cortical localization of other polarity factors, including and that tea1p and mod5p can also interact directly, the actin-binding protein bud6/aip3 (Glynn et al, 2001; independent of tea3p. Despite their related structures, Jin and Amberg, 2001), the formin for3p (Feierbach and different regions of tea1p and tea3p are required for their Chang, 2001; Feierbach et al, 2004), the SH3-domain protein respective interactions with an essential central region of tea4p/wsh3p (Martin et al, 2005; Tatebe et al, 2005) and the mod5p. We demonstrate that tea3p is required for proper tea1p-related protein tea3p (Arellano et al, 2002). cortical localization of tea1p, specifically at nongrowing Tea1p is a 1147-amino-acid protein. Its N-terminus con- cell tips, and that tea1p and mod5p are independently tains six kelch repeats, which are found in many proteins, required for tea3p localization. Further, we find that tea3p including actin-binding proteins, and are likely involved in fused to GFP or mCherry is cotransported with tea1p by protein–protein interactions (Adams et al, 2000; Prag and microtubules to cell tips, but this occurs only in the Adams, 2003; Li et al, 2004). The C-terminal regions of tea1p absence of mod5p. These results suggest that independent are predicted to be largely alpha-helical coiled coil. Tea1p also protein–protein interactions among tea1p, tea3p and displays limited functional and sequence similarity to ezrin, mod5p collectively contribute to tea1p anchoring at cell a member of the ERM (ezrin–radixin–moesin) family, which tips via a multistep and multimode mechanism. has been shown to link the plasma membrane with the The EMBO Journal (2005) 24, 3690–3699. doi:10.1038/ underlying actin cytoskeleton in animal cells (Vega and sj.emboj.7600838; Published online 13 October 2005 Solomon, 1997; Bretscher et al, 2002). Subject Categories: cell & tissue architecture In vivo, tea1p displays a dynamic pattern of localization, Keywords: cell polarity; kelch repeat; mCherry; S. pombe; reflecting its function in cell polarity. Tea1p is transported to tdTomato cell tips on the growing plus ends of microtubules (Behrens and Nurse, 2002; Snaith and Sawin, 2003; Feierbach et al, 2004), dependent on the kip2-like kinesin tea2p and the CLIP- 170 homologue tip1p (Browning et al, 2000, 2003; Brunner Introduction and Nurse, 2000; Busch et al, 2004). Once at tips, tea1p is unloaded from microtubules and retained at the cell cortex. Communication between microtubules and the actin cyto- This retention is defective in mutants lacking either the skeleton is a common feature of eukaryotic cell polarity membrane protein mod5p or the C-terminal 200 amino (Drubin and Nelson, 1996; Yarm et al, 2001; Chang and acids of tea1p (tea1D200; Behrens and Nurse, 2002; Martin Peter, 2003; Small and Kaverina, 2003). In many cases, and Chang, 2003; Snaith and Sawin, 2003; also see below). proteins that are transported on the plus ends of micro- Cortically associated tea1p then contributes to the organiza- tubules, such as EB1 and CLIP-170, forge interactions with tion of actin filaments at cell tips via interactions with proteins proteins localized at sites on the plasma membrane, facilitat- such as tea4p/wsh3p and bud6p/aip3p, thereby ensuring ing this communication (Schuyler and Pellman, 2001; proper bipolar growth (Glynn et al, 2001; Verde, 2001; Gundersen, 2002; Carvalho et al, 2003; Akhmanova and Martin et al, 2005; Snaith and Sawin, 2005; Tatebe et al, 2005). Hoogenraad, 2005; Watanabe et al, 2005). Currently, there are three major outstanding issues in In the fission yeast Schizosaccharomyces pombe, tea1p is relation to how tea1p functions in microtubule-mediated a key mediator between microtubules and the actin cyto- cell polarity in fission yeast. At the beginning of the tea1p ‘pathway’ is the question of how tea1p is associated with and *Correspondence: Wellcome Trust Centre for Cell Biology, Swann Building, School of Biological Sciences, University of Edinburgh, transported on microtubule plus ends, and at the end of the Mayfield Road, Edinburgh EH9 3JR, UK. Tel.: þ 44 131 650 7064; pathway is the question of how tea1p interacts with the actin Fax: þ 44 131 650 7360; E-mail: [email protected] cytoskeleton. Here, we address a question at the center of the tea1p pathway, namely, how does tea1p become anchored at Received: 21 March 2005; accepted: 15 September 2005; published online: 13 October 2005 the cortex at cell tips? 3690 The EMBO Journal VOL 24 NO 21 2005 &2005 European Molecular Biology Organization | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al An important factor in tea1p anchoring is the protein A α-GST IP WCE mod5p, which was first identified in a screen for mutations + Δ + + + Δ + + GST-mod5 in S. pombe that affect cell shape (Snaith and Sawin, 2003). tea3-GFP + + Δ + + + Δ + In mod5D cells, tea1p is transported normally to cell tips but + + + Δ + + + Δ tea1 fails to accumulate there (Snaith and Sawin, 2003). Mod5p, GST-mod5p which is unusually rich in serine, threonine and proline residues and thus may be an intrinsically unstructured pro- tea3p-GFP tein (Dyson and Wright, 2005), is localized to membranes at tea1p cell tips, via a C-terminal prenylation sequence. Interestingly, in tea1D cells, mod5p is no longer restricted to cell tips but 13 2 4 5 6 7 8 rather spreads out around the entire cell cortex. This has led α-HA IP WCE to the suggestion of a positive feedback loop linking mod5p B mod5 + + Δ + + Δ and tea1p localization, which might ensure high-fidelity tea3-HA – + + – + + polarized growth in the context of dynamic tea1p delivery tea1 + + + + + + to cell tips (Snaith and Sawin, 2003). However, the molecular mechanisms underlying tea1p localization at the cell cortex tea3p-HA remain largely unclear. For example, although there is genetic evidence for an interaction between mod5p and tea1p, a clear tea1p biochemical interaction has not yet been shown. In particu- 1324 5 6 lar, live-cell analyses of GFP-tagged mod5p and tea1p have demonstrated that the two proteins do not completely co- C kelch repeats Coiled coil localize at cell tips, and that tea1p at the cortex is highly tea1p 1 1147 78 370 609 1106 dynamic (Snaith and Sawin, 2003). Thus, even if tea1p and mod5p Membrane association/ binding tea3p binding mod5p do interact physically, whether directly or indirectly, Coiled coil other proteins may also be involved in regulating their kelch repeats tea3p 1 1125 localization. 179 371 599 966 In this work, we have sought to further understand the mod5p tea1p binding binding mechanisms regulating tea1p localization at cell tips by Prenylation identifying protein–protein interactions involving mod5p. signal From a two-hybrid screen, we identified the tea1p-related mod5p 1 522 protein tea3p as a mod5-interacting protein. In these experi- tea1p/tea3p binding ments, we also found that mod5p interacts with tea1p and that tea1p interacts with tea3p. In vivo, each of the pairwise WCE GST-IP tea1p-IP interactions occurring between any two of these three pro- tea1 + Δ200 + Δ200 + Δ200 teins can take place in the absence of the third protein, and in a series of localization-dependency studies, we found that GST-mod5 each of these proteins contributes to the proper localization tea1p tea1Δ200p and function of the other two, but in distinct ways. These results suggest that a complex network of interactions among tea3p-HA tea1p, mod5p and tea3p is involved in regulating the locali- 13 2 4 5 6 zation of these three proteins, ultimately leading to correctly Figure 1 Tea1p, mod5p and tea3p form independent complexes anchored tea1p at cell tips. in vivo.(A) GST-mod5p was immunoprecipitated from soluble protein extracts of wild-type cells expressing GST-mod5p and tea3p-GFP (lanes 1 and 5), mod5D cells expressing tea3p-GFP (lanes 2 and 6), Results tea3D cells expressing GST-mod5p (lanes 3 and 7) and tea1D cells expressing GST-mod5p and tea3p-GFP (lanes 4 and 8). The resulting Pairwise interactions between tea1p, tea3p and mod5p immunocomplexes were analyzed for GST-mod5p, tea3p-GFP and To better understand the mechanisms underlying tea1p loca- tea1p. Whole-cell extract (WCE) fractions are shown in lanes 1–4 lization. we performed a two-hybrid screen using essentially and immunoprecipitates in lanes 5–8. Immunoprecipitates were full-length mod5p as bait (see Supplementary data). Out of 11 loaded 30 relative to WCE sample. (B) Tea3p-HA was immuno- precipitated from soluble protein extracts of wild-type cells (lanes 1 plasmids isolated, 10 contained fragments of the tea3 gene, and 4), wild-type cells expressing tea3p-HA (lanes 2 and 5) and previously implicated in cell polarity (Arellano et al, 2002). mod5D cells expressing tea3p-HA (lanes 3 and 6). The resulting Interestingly, tea3p is structurally related to tea1p. Both immunocomplexes were analyzed for tea3p-HA and tea1p. WCE proteins contain the protein–protein interaction kelch domain fractions are shown in lanes 1–3 and immunoprecipitates in lanes 4–6. Immunoprecipitates were loaded 20 relative to WCE. in their N-termini (Adams et al, 2000; Prag and Adams, 2003; (C) Schematic diagram summarizing interactions between tea1p, tea3p Li et al, 2004) and long regions of predicted coiled coil in and mod5p. (D) GST-mod5p (lanes 3 and 4) or tea1p (lanes 5 and 6) their C-termini. We confirmed the interaction between mod5p was immunoprecipitated from soluble protein extracts of either and tea3p in GST pulldown experiments and co-immuno- wild-type (lanes 1, 3 and 5) or tea1D200 (lanes 2, 4 and 6) cells expressing GST-mod5p and tea3p-HA. The resulting immunocom- precipitation assays, using extracts from fission yeast plexes were analyzed for GST-mod5p, tea1p (or tea1D200p) and simultaneously expressing GST-mod5p and tea3p-GFP tea3p-HA. WCE fractions are shown in lanes 1 and 2 and immuno- (Supplementary Figure 1A and Figure 1A). precipitates in lanes 3–6. Immunoprecipitates were loaded 30 Arellano et al (2002) reported that tea3p bound to tea1p relative to WCE. in the yeast two-hybrid system, but they did not verify this &2005 European Molecular Biology Organization The EMBO Journal VOL 24 NO 21 2005 3691 | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al biochemically. We found that immunoprecipitation of tea3p- Full-length or mutant GFP-mod5p expression in: HA from fission yeast cell extracts co-precipitated tea1p mod5Δ background mod5 + bkgd (Figure 1B). In addition, we found that tea1p was co- tea1p/tubulin GFP signal tea1p/tubulin precipitated in immunoprecipitates from cells expressing A EI GST-mod5p (Figure 1A). Thus, tea1p, tea3p and mod5p all associate with each other in vivo. Based on these results, we wanted to examine whether tea3p might act as a molecular ‘bridge’ between tea1p and mod5p. We therefore used deletion strains to test whether GFP-mod5p full-length each of the pairwise protein–protein interactions observed in immunoprecipitation experiments could occur in the absence B F J of the third protein. As expected from two-hybrid results, tea1p interacted with tea3p in the absence of mod5p, and mod5p interacted with tea3p in the absence of tea1p (Figure 1A and B). Interestingly, however, we also found that tea1p interacted with mod5p in the absence of tea3p (Figure 1A). GFP-mod5Δ256 – 305p Thus, each of the pairwise interactions among tea1p, tea3p and mod5p is independent of the third protein. C G K Through a combination of immunoprecipitation experi- ments, two-hybrid analysis and in vitro binding studies, we mapped the regions of tea1p, tea3p and mod5p involved in binding to each other (Supplementary Figures 1B–F, 2A and B). The results are summarized in Figure 1C. Four important GFP-mod5Δ156 – 205p points emerged from these experiments. First, all of the observed interactions are likely to be direct. Second, a central D H L region of mod5p (amino acids 156–205) is required for binding to both tea1p and tea3p. Third, even though tea1p and tea3p are structurally related, binding of tea1p to mod5p is mediated by the N-terminus of tea1p (amino acids 1–352), while binding of tea3p to mod5p is mediated by a central GFP-mod5Δ206 – 255p coiled-coil region of tea3p (amino acids 739–785). Fourth, binding of tea1p to tea3p is mediated by the C-termini of Figure 2 Mod5p amino acids 156–255 are essential for the localiza- both proteins (amino acids 948–1147 of tea1p and amino tion of tea1p and mod5p. The localization of tea1p (green) and microtubules (red) in (A–D) mod5D cells and (I–L) wild-type cells, acids 901–1125 of tea3p). These last two points are particu- expressing different mutant versions of GFP-mod5p. (E–H) The larly salient because Behrens and Nurse (2002) have shown localization of GFP-mod5p (and mutant versions) in mod5D cells. that deletion of the tea1p C-terminus (tea1D200) prevents (A, E, I) Wild-type GFP-mod5p; (B, F, J) GFP-mod5D256–305p; anchoring of tea1p to the cell cortex. In this context, our (C, G, K) GFP-mod5D156–205p; (D, H, L) GFP-mod5D206–255p. The scale bar represents 5mm. mapping data suggest that the anchoring defect of tea1D200 mutants correlates not with a failure to bind mod5p but rather with a failure to bind tea3p and/or other proteins (see Discussion). The central region of mod5p is required for function We next tested whether tea1p, tea3p and mod5p all coexist We next wanted to determine to what extent the protein– in a single protein complex in vivo. This was not entirely protein interactions identified might mediate the localization straightforward to determine, because physical analysis of and/or function of mod5p, tea3p and tea1p, in order to tea1p shows it to be present in complexes covering a wide understand how these interactions contribute to the proper range of molecular sizes (Feierbach et al, 2004). Moreover, cortical anchoring of tea1p at cell tips. because each of the three proteins concerned can interact The only recognizable amino-acid sequence motif in independently with the other two, immunoprecipitation of mod5p is a C-terminal prenylation signal that is essential for any one protein will co-precipitate both of the other two, even mod5p function and localization (Snaith and Sawin, 2003). To if a three-way complex does not exist. We therefore examined identify other functionally significant regions, we fused GFP to whether a three-way complex could be identified when one a series of 50-amino-acid internal deletions spanning the of the pairwise interactions was disrupted. Interestingly, we mod5p open reading frame (i.e., those used in mapping found that tea3p could not be co-immunoprecipitated with studies; Supplementary Figure 2A) and expressed the inter- tea1D200p, even though both proteins could bind to mod5p nal-deletion mutant proteins individually in mod5D cells. In a in the same cell extract (Figure 1D). This suggests that a quantitative polarity maintenance assay, nearly all of the three-way complex of tea1p, tea3p and mod5p may be deletions behaved like wild-type mod5p; only cells expressing present only transiently in vivo, if at all. However, we also either mod5D156–205p or mod5D206–255p, which fail to bind found that the absence of a three-way complex in vivo is tea1p or tea3p (Supplementary Figure 2A and B), were unable unlikely to be a result of tea1p and tea3p competing for to maintain polarity (Supplementary Figure 3A). potentially overlapping binding sites on mod5p, as a three- We also studied the localization of tea1p and the GFP- way complex could be demonstrated artificially, in a yeast mod5p internal-deletion proteins themselves in each internal- ‘bridging two-hybrid’ assay (Supplementary Figure 2C). deletion mutant strain. In nearly all cases, both tea1p and the 3692 The EMBO Journal VOL 24 NO 21 2005 &2005 European Molecular Biology Organization | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al mutant versions of mod5p were localized as in wild-type cells tea3p-GFP (Figure 2A, B, E and F; additional data not shown). How- A B C ever, in cells expressing either mutant mod5D156–205p or mod5D206–255p, tea1p localization resembled that of mod5D cells, with tea1p present on microtubule plus ends but failing to accumulate at the cell cortex (Figure 2C and D and Supplementary Figure 4E). These same two mutants were also defective in mod5 localization (Figure 2G and H), Wild type tea1ΔΔ tea1 200 with the mutant GFP-mod5p spread out around the plasma D E F membrane in the same manner as full-length wild-type mod5p in tea1D cells (Snaith and Sawin, 2003). Tea1p localization was not altered from wild type when any of the internal-deletion proteins was expressed in mod5 rather than mod5D backgrounds, even though the localization of GFP-tagged mod5D156–205p and mod5D206–255p remained mod5ΔΔ mod5 tea1ΔΔ mod5 206 – 255 aberrant in these experiments (Figure 2I–L; additional data Figure 3 Localization dependencies of tea3p-GFP. Localization of not shown). These results suggest that, in addition to the tea3p-GFP in (A) wild-type, (B) tea1D,(C) tea1D200,(D) mod5D, C-terminal prenylation site, the central region of mod5p, (E) mod5D tea1D and (F) mod5D206–255 cells. The scale bar represents 5mm. which interacts with both tea1p and tea3p, contains the major determinants for both mod5p localization and function. We previously showed that restriction of mod5p to cell tips is dependent on tea1p and also partially on tea3p notion that tea1p can promote the localization of tea3p to cell (Supplementary Figure 3B and C; Snaith and Sawin, 2003). tips, while mod5p acts to stably integrate tea3p at the cell In light of our results showing that mod5p can bind to cortex, to the extent that the mislocalization of mod5p (i.e., tea1D200p, we examined the localization of GFP-mod5p in in tea1D mutants) will recruit tea3p to ectopic cortical sites. tea1D200 cells. GFP-mod5p was spread out around the We also found that in cells expressing mod5D156–205p membrane (Supplementary Figure 3D), suggesting that or mod5D206–255p, tea3p-GFP was delocalized in a pattern restriction of mod5p to cell tips requires not only the similar to that present in mod5D cells (Figure 3F; data not tea1p–mod5p interaction but also the stable binding of shown). In conclusion, the overall tip localization of tea3p is tea1p at the cell cortex. dependent both on binding to mod5p and on the carboxy- terminus of tea1p. Tea1p and mod5p are independently required for different aspects of tea3p localization We next sought to investigate the roles played by tea1p and Microtubule-based transport of tea3p to cell tips mod5p in the localization of tea3p. In wild-type cells, the requires both tea1p and the absence of mod5p majority of tea3p-GFP was confined to the cell tip region, Because tea1p binds to tea3p and facilitates its localization to with a few cytoplasmic dots also present, confirming pre- cell tips, and tea1p is transported to cell tips by association vious results (Figure 3A; Arellano et al, 2002). In contrast, in with microtubule plus ends (Behrens and Nurse, 2002; Snaith tea1D cells, tea3p-GFP was still partially enriched at cell tips and Sawin, 2003; Feierbach et al, 2004), we investigated but also displayed a punctate staining evenly spread around whether tea3p is similarly transported to cell tips along the plasma membrane (Figure 3B; Arellano et al, 2002), microtubules. In wild-type cells, tea3p-GFP displayed little reminiscent of GFP-mod5p localization in tea1D mutants or no apparent movement toward cell tips (Supplementary (Snaith and Sawin, 2003). In addition, we observed an Movie 1). Previously, however, we demonstrated that levels increase in the cytoplasmic tea3p-GFP signal (Figure 3B). In of tea1p-GFP transport to cell tips are elevated in mod5D cells tea1D200 cells, the same mislocalization of tea3p-GFP was relative to wild-type cells, and we argued that this might be seen (Figure 3C), consistent with the carboxy-terminal region due to the fact that an inability of tea1p to anchor at the of tea1p being required for binding to tea3p (Supplementary cortex could result in a higher free cytoplasmic pool of tea1p Figure 1F). able to associate with microtubules (Snaith and Sawin, 2003). Because mod5p is required for proper tea1p localization, We therefore followed tea3p-GFP in mod5D cells. and tea1p is required for tea3p localization, we suspected that Here, we observed significant motion of cytoplasmic tea3p- mod5p would also be necessary for the correct localization GFP particles. Most of the fainter particles of tea3p-GFP of tea3p. Interestingly, however, the manner of tea3p-GFP showed rapid movements with no clear directionality. mislocalization in mod5D mutants was different from that in However, many of the bright particles of tea3p-GFP exhibited tea1D mutants. In mod5D cells, the tea3p-GFP signal became linear movements, both toward and away from cell tips highly punctate, with a few bright dots of tea3p-GFP visible (Figure 4A and Supplementary Movie 2). To test whether near cell tips and also throughout the cell, suggesting that this movement was dependent on the actin or microtubule mod5p is required for a stable membrane localization cytoskeletons, we treated cells with latrunculin B or methyl of tea3p (Figure 3D). Consistent with this, we found that 2-benzimidazolecarbamate (carbendazim; MBC). Treatment tea3p-GFP localization in mod5D tea1D double mutants of cells with latrunculin B for up to an hour had no significant resembled tea3p-GFP in mod5D single mutants, although effect on the movement of tea3p-GFP (data not shown). By tea3p-GFP localization to cell tips was more compromised contrast, within 5 min of MBC treatment, the movement of than in single mutants (Figure 3E). These results support the bright tea3p-GFP particles was reduced, and after 20 min it &2005 European Molecular Biology Organization The EMBO Journal VOL 24 NO 21 2005 3693 | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al mod5Δ mod5Δ Wild type tea3p-GFP GFP-atb2 /tea3p-mCh GFP-atb2 /tea3p-mCh no MBC plus MBC A B C Wild type tea1p-GFPt /tea3p-mCh Wild type 88 – 12 – mod5Δ 57 37 6 – tea3Δ 15 77 5 – tea1Δ 12 76 – 8 Figure 5 Mod5p and tea3p have distinct but overlapping functions. (A) Wild-type, mod5D, tea3D, tea1D, mod5D tea3D, mod5D tea1D, mod5ΔΔ mod5 tea1Δ mod5Δ tea1p-GFP/tea3p-mCh tea3p-GFP tea3D tea1D and tea1D200 cells were depolarized by growth to tea1p-GFP/ tea3p-mCh stationary phase and returned to fresh medium in the absence (gray E F G bars) or presence (black bars) of 50mg/ml MBC for 3 h at 321C. The percentage of branched cells in each sample was counted, n¼ 200. (B) Percentage of daughter cell pairs displaying illustrated initial growth patterns after septation in wild-type (n¼ 133), mod5D (n¼ 228), tea3D (n¼ 165) and tea1D cells (n¼ 194). Arrows indicate direction of growth. conjunction with GFP-tubulin (GFP-atb2p; Adachi et al, 1986). In wild-type cells, tea3p-mCh accumulated at cell tips, consistent with the localization of tea3p-GFP, although in further experiments we found that fusion of tea3p with mCherry may in fact compromise some more subtle aspects of tea3p behavior as compared to fusion with GFP (data not shown). When we imaged tea3p-mCh together with GFP- tubulin in mod5D mutants, tea3p-mCh particles were readily detected on the ends of growing microtubules (Figure 4B). In wild-type cells, however, tea3p-mCh particles failed to colocalize with microtubule ends (Figure 4C). Figure 4 Microtubule-dependent movement of tea3p in mod5D. We next used tea3p-mCh to test whether tea3p was co- (A) Time-lapse movie frames of tea3p-GFP in mod5D cells at 15 s intervals. Red and turquoise arrowheads mark traveling particles of transported with tea1p to cell tips. In wild-type cells, some tea3p-GFP. (B) Time-lapse movie frames of GFP-atb2p (green) and cortical particles of tea3p-mCh colocalized with tea1p-GFP, tea3p-mCh (red) in mod5D cells at 15 s intervals. Gray dashed line while others remained distinct from tea1p-GFP (Figure 4D), indicates the starting position of the traveling tea3p-mCh particle. (C) Localization of GFP-atb2p (green) with tea3p-mCh (red) in wild- suggesting that there may be distinct populations of each type cells. (D) Localization of tea1p-GFP (green) with tea3p-mCh protein, with an overlapping subset. In mod5D cells, particles (red) in wild-type cells. (E, F) Time-lapse movie frames of tea1p- of tea3p-mCh colocalized with tea1p-GFP as they translocated GFP (green) and tea3-mCh (red) in mod5D cells at 15 intervals. both toward and away from the cell tips (Figure 4E and F). Traveling particles of colocalized tea1p-GFP and tea3p-mCh are (As described in Supplementary data, relatively long expo- indicated by white arrowheads and static particles are indicated by white arrows. (G) Time-lapse movie frames of tea3p-GFP in sure times were required during time-lapse acquisition, re- mod5 tea1D cells at 15 s intervals. Red dotted lines indicate reduced sulting in occasional misalignment of the moving tea3p-mCh movement of tea3p-GFP. The scale bar represents 5mm. and tea1p-GFP signals.) Thus, tea3p and tea1p can move together. Finally, we tested whether the movement of tea3p-GFP was almost abolished (data not shown), suggesting that tea3- seen in mod5D cells was dependent on tea1p. In mod5D GFP was moving on microtubules. tea1D double mutant cells, the linear movement of bright To confirm this, we developed a series of tagging plasmids particles of tea3p-GFP was almost completely abolished derived from novel variants of the red fluorescent protein (Figure 4G and Supplementary Movie 3). Collectively, these dsRed (Shaner et al, 2004; see Supplementary data) and used results suggest that tea3p does not significantly associate these to construct strains carrying tea3p tagged with the with tea1p or microtubules in the cytoplasm in wild-type variant mCherry (tea3p-mCh), in order to image tea3p in cells. However, in mod5D cells, where tea3p is no longer 3694 The EMBO Journal VOL 24 NO 21 2005 &2005 European Molecular Biology Organization | | % branched cells Wild type mod5Δ tea3Δ tea1 mod5Δ tea3Δ mod5Δ tea1 tea3Δ tea1Δ tea1Δ200 Tea1p, mod5p and tea3p in fission yeast HA Snaith et al sequestered at cells tips, tea3p can associate with tea1p in the delivery of tea1p to cell tips (Snaith and Sawin, 2003). We cytoplasm and as a result is cotransported with tea1p along found that tea3D cells, like mod5D cells, formed branches microtubules to cell tips. only in the presence of MBC (Figure 5A). If tea3p and mod5p both acted in a strictly linear genetic pathway to regulate cell Mod5p and tea3p have distinct but overlapping polarity, we might expect that the phenotype of the mod5D functions in cell polarity and tea1p anchoring tea3D double mutant would resemble the single mutants. Both mod5D and tea3D mutants have been shown to have However, a significant fraction of mod5D tea3D mutants growth-polarity defects, but thus far each mutant has been (20%) formed branched cells even without microtubule dis- examined in a distinct type of polarity assay (Arellano et al, ruption, much more than occurs in either single mutant 2002; Snaith and Sawin, 2003). To determine whether mod5p (Figure 5A). This suggests that although mod5p and tea3p and tea3p act in a single common pathway, we tested several both interact and contribute to growth polarity in fission mutants side by side in both assays. yeast, each protein may make a distinct contribution to cell In return-to-growth experiments involving cells previously polarity and tea1p function. grown to stationary phase, wild-type cells re-establish polar- In these experiments, we also found that tea1D200 cells ity axes at pre-existing cell ends, and tea1D mutants form formed branches both in the presence and absence of MBC branched cells (Browning et al, 2000; Snaith and Sawin, (Figure 5A) and thus resemble tea1D rather than mod5D 2003). By contrast, mod5D mutants re-establish polarity or tea3D mutants with regard to polarity defects. Because normally when microtubules are intact but form branches tea1D200p can still bind to mod5p (Supplementary Figure when microtubules are disrupted with MBC, which impairs 1E), this suggests that the essential function of the C-terminus A B C – MBC Wild type tea3 Δ orb2-34 D E F + MBC Wild type tea3 Δ orb2-34 tea3Δ G H Wild type 01 5 0 15 20 01 5 0 15 20 Time in MBC (min) Time in MBC (min) neither end 1 end 2 ends neither end 1 end 2 ends I J K – MBC L M N + MBC Figure 6 Tea1p is preferentially lost from the nongrowing tip in tea3D cells. (A–H) Treatment of (A, D) wild-type, (B, E) tea3D and (C, F) orb2- 34 cells with MBC for (A–C) 0 min or (D–F) 5 min. Cells are stained for tea1p (green) and microtubules (red). Time course of tea1p loss from cell tips in (G) wild-type and (H) tea3D cells, showing number of cells with detectable levels of tea1p at one cell tip, two tips or neither tip, for each strain. n¼ 200 for each strain. (I–N) Treatment of tea3D cells with MBC for (I–K) 0 min or (L–N) 5 min. Cells are stained with anti-tea1p antibodies (I, L) and Alexa-labeled phalloidin (J, M). Merged imaged are shown with tea1p in green and phalloidin in red (K, N). The scale bars represent 5mm. &2005 European Molecular Biology Organization The EMBO Journal VOL 24 NO 21 2005 3695 | | Number of cells Number of cells Tea1p, mod5p and tea3p in fission yeast HA Snaith et al of tea1p in anchoring tea1p may be required subsequent to symmetrically from both cell tips (Figure 6A and D), tea3D the tea1p–mod5p interaction. In addition, because tea1D200p cells rapidly lost tea1p from one tip, such that after 5 min of fails to bind tea3p but the tea1D200 phenotype is more MBC treatment, nearly all tea3D cells displayed monopolar extreme than the tea3D phenotype, this indicates that the tea1p localization (Figure 6B and E). To determine whether tea1p carboxy-terminus is likely to interact functionally not this asymmetric loss of tea1p was simply due to the mono- only with tea3p but also with other proteins as well (see polar growth pattern of tea3D cells, we assayed tea1p loca- Discussion). lization in the pak1/shk1/orb2 mutant orb2-34, which also The second assay involved time-lapse observations of exhibits strongly monopolar growth (Verde et al, 1995; Kim growth patterns in the different mutants. Whereas tea3D et al, 2003). In contrast to tea3D cells, orb2-34 cells lost tea1p cells mainly exhibit monopolar growth, in a pattern similar symmetrically after treatment with MBC (Figure 6C and F). to tea1D mutants (Glynn et al, 2001; Arellano et al, 2002), we We also examined actin and tea1p localization in MBC- found that the majority of mod5D cells display wild-type treated tea3D cells to determine whether tea1p was lost growth patterns, with less than 50% of cells showing the from the growing tip, the nongrowing tip or randomly from aberrant growth patterns characteristic of tea1D and tea3D either tip (Figure 6I–N). In nearly all tea3D cells, tea1p was mutants (Figure 5B). This result strongly supports a view lost from the nongrowing, actin-poor tip; the small number of of overlapping but distinct functions for mod5p and tea3p. cases in which loss of tea1p was symmetrical (7%, n¼ 100) Moreover, since the monopolar growth defects in mod5D corresponded to those few tea3D cells that had initiated mutants are much less severe than in tea1D mutants, these bipolar growth. These results indicate that tea3p is required data also suggest that at least some ‘bipolar growth’ functions for the proper cortical anchoring of tea1p, specifically at of tea1p can bypass the need for mod5p at cell tips (see nongrowing cell tips. Discussion). While we have shown that mod5p is required for anchor- Discussion ing of tea1p at cell tips (Snaith and Sawin, 2003), Arellano et al (2002) reported that tea3p does not play a role in tea1p Multistep and multimode tea1p anchoring localization. Our studies with these proteins prompted us to We have demonstrated multiple interactions among tea1p, re-examine the localization of tea1p in tea3D mutants. We tea3p and mod5p and examined their roles in cell polarity found that although tea1p is present at both tips of tea3D and in regulating protein localization at cell tips, especially cells, tea3D cells have only 76% of the tea1p levels seen at that of tea1p. Because interactions between any two of these cell tips in wild-type cells, while tea1p localization in mod5D three proteins can occur in vivo in the absence of the third tea3D double mutants resembles tea1p localization in mod5D protein, while mutation and/or deletion of any one protein cells (Supplementary Figure 4). Because this reduction of tip- characteristically alters the localization of each of the other localized tea1p in tea3D mutants could reflect a role for tea3p two, it might appear difficult to place these interactions at in anchoring tea1p at cell tips, we examined tea1p levels at specific points in a pathway regulating tea1p anchoring. cell tips in wild-type cells and tea3D mutants after micro- However, by taking into account our additional findings— tubule disruption by MBC, where new tea1p delivery to cell for example, the asymmetric cell-tip loss of tea1p in tea3D tips is prevented. Under these conditions, tea1p is rapidly lost mutants, the cotransport of tea3p with tea1p only in the from both cell tips in mod5D mutants (HA Snaith, unpub- absence of mod5p and the behavior of tea1D200p—we have lished data). Whereas wild-type cells gradually lost tea1p formulated a multistep and multimode model for the efficient Nongrowing cell tip tea1p tea3p C tea3p mod5p anchors tea1p and tea3p tea3p at cortex 'bypass' mod5p mod5p tea1p tea1p interacts (mod5Δ) with tea3p tea3p A B tea1p tea1p tea1p tea1p delivered by mod5p MTs to cortex tea1p interacts tea1p localization at with mod5p at both cell tips both cell tips tea1p tea1p interacts with X (= tea4p?) Growing cell tip Figure 7 Multistep cortical retention of tea1p. A model of protein–protein interactions regulating tea1p localization at cell tips. (A) Tea1p arrives at the cell cortex on microtubule plus ends. (B) The tea1p N-terminus interacts with mod5p at both cell tips. (C) At nongrowing cell tips, the tea1p C-terminus interacts with tea3p, which is itself anchored at the cortex (D) via interaction with mod5p. (E) At growing cell tips, tea1p interacts with unknown factors (X) that provide a function similar to that of tea3p at nongrowing tips. (F) A partially functional interaction between tea1p and tea3p can occur without mod5p, but tea1p and tea3p will be poorly anchored. (G) Together, these interactions lead to proper tea1p anchoring at both cell tips. See text for further details. 3696 The EMBO Journal VOL 24 NO 21 2005 &2005 European Molecular Biology Organization | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al anchoring of tea1p at cell tips (Figure 7). The model is interact with tea3p, we have placed a tea1p–tea3p interaction ‘multistep’ because it postulates that efficient anchoring of subsequent to the tea1p–mod5p interaction at nongrowing tea1p at the cortex occurs in several stages, and ‘multimode’ cell tips, which accords with the demonstrated role for tea3p because we propose that the mechanisms of tea1p anchoring in the retention of tea1p at these sites. Furthermore, since are different in the presence or absence of mod5p, and at deletion of tea3p does not phenocopy tea1D200, we propose growing versus nongrowing cell tips. According to the model, that there may also be a ‘parallel mode’ of tea3p-independent tea1p is initially in a transport complex on microtubule plus tea1p retention at the growing cell tip (see below). ends, in association with other plus-end binding proteins (Figure 7A; also see below; Busch et al, 2004; Feierbach Parallel tea1p-anchoring systems at growing and nongrowing et al, 2004). Upon reaching cell tips, tea1p interacts, via its tips. Our results suggest that there are both common and N-terminus, with a central region of mod5p (Figure 7B). This distinct elements in the mechanisms regulating tea1p locali- interaction is important for the anchoring of tea1p at the zation at growing versus nongrowing cell tips. Whereas cortex (Snaith and Sawin, 2003) but is by itself insufficient mod5p and the tea1p C-terminus are each required for for the correct steady-state localization of tea1p and mod5p. tea1p anchoring at both cell tips, tea3p is required primarily Tea1p then further interacts, via its C-terminus, with corti- for tea1p anchoring at nongrowing tips, consistent with its cally localized tea3p. This may occur at both cell tips but preferential localization to nongrowing tips (Arellano et al, is functionally important primarily at nongrowing tips 2002). Thus, at growing cell tips, other tea1p-binding proteins (Figure 7C). Our data further indicate that tea3p is itself may perform functions analogous to those performed by anchored at cell tips by a strong interaction with a central tea3p at nongrowing tips. Further support for the suggestion region of mod5p (Figure 7D). We also suggest that at growing that a parallel pathway at growing cell tips involves the cell tips, there may be a parallel, tea3p-independent pathway tea1p–mod5p interaction comes from the observation that of tea1p retention, involving both a tea1p–mod5p interaction mod5D tea3D double mutants form ectopic branches at a and an interaction of tea1p with other cortical proteins much higher frequency than tea3D single mutants (Figure 7E). Finally, we propose that in addition to the (Figure 3A). This indicates that mod5p makes additional mod5p-dependent pathway of cortical tea1p anchoring, a contributions to polarity fidelity outside a strictly ‘tea3p- complementary mechanism exists whereby tea1p and tea3p specific’ pathway. Recently, a novel SH3-domain protein, functionally interact both in the cytoplasm and at the cortex tea4p/wsh3p, has been identified that may perform the to ‘bypass’ mod5p (Figure 7F). Tea1p anchored at cell tips ‘tea3p-analogous’ function at growing cell tips; interestingly, (Figure 7G) would then be capable of making further inter- tea4p, like tea3p, binds to the C-terminus of tea1p (Martin actions with its ‘downstream’ effectors, perhaps also further et al, 2005; Tatebe et al, 2005). strengthening its retention as a result. The tea1p–tea3p interaction can ‘bypass’ mod5p. This sug- Structure of the model gestion is based not only on the strong tea1p–tea3p inter- action seen in mod5D cells but also on mutant cell-polarity Because of the diverse nature of the results contributing to phenotypes. Both tea1D and tea3D mutants have more severe our model, we discuss more specific aspects of the model monopolar growth defects than mod5D (Figure 5B), indicat- below. ing that tea1p and tea3p still retain some function in organiz- A cortical interaction between tea3p and mod5p. Several ing growth polarity in the absence of mod5p, even if they are pieces of evidence support the idea that tea3p and mod5p not stably anchored at cell tips during steady-state growth. form a stable cortical complex. First, tea3p and mod5p have a Further support comes from the observation that mod5D tea3D double mutants form ectopic branches at a much robust biochemical interaction both in the presence and higher frequency than mod5D single mutants (Figure 5A). absence of tea1p; indeed, tea3p was the only strong interactor Moreover, antipodal growth persists in mod5D mutants as with mod5p identified in our two-hybrid screen. Second, we long as there is constant delivery of tea1p to cell tips by note that the delocalized tea3p seen in tea1D cells is never- theless still associated with the cell cortex, in a pattern microtubules (Snaith and Sawin, 2003). Most importantly, closely resembling that of mod5p (Figure 3; Arellano et al, our proposal for a ‘mod5p bypass’ is directly supported by the observation that tea3p is cotransported with tea1p to cell 2002). As this cortical localization is lost when both tea1p tips when mod5þ is deleted (Figure 4). and mod5p are deleted, it seems likely that the aberrant cortical localization of tea3p in tea1D mutants is due to its Limits of the model. An important issue not explicitly stable interaction with mod5p. addressed in the model is how the hypothesized positive The C-terminus of tea1p is important subsequent to the tea1p– feedback between tea1p and mod5p localization may operate (Snaith and Sawin, 2003). We have shown that two mod5 mod5p interaction. Tea1D200p is found at microtubule plus internal-deletion mutants fail to be restricted to cell tips even ends at both cell tips but is not properly anchored at the when tea1p is properly localized at the cortex, but we cannot cortex (Behrens and Nurse, 2002). We had initially expected yet distinguish whether this is directly due to their failure to this to be due to a failure of tea1D200p to bind to membrane- associated mod5p, but instead we found that tea1D200p can bind to tea1p, or to tea3p, or both. This will benefit from interact with mod5p in vivo. This suggests that the ‘failure further studies. point’ of tea1D200p in a tea1p-anchoring pathway occurs after the tea1p–mod5p interaction, and that interactions of Concluding remarks the tea1p C-terminus with other proteins are additionally In summary, a complex set of pairwise protein–protein inter- required for tea1p anchoring. Because tea1D200p fails to actions among tea1p, tea3p and mod5p is essential for the &2005 European Molecular Biology Organization The EMBO Journal VOL 24 NO 21 2005 3697 | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al Immunoprecipitation studies correct cortical anchoring of tea1p after its delivery to cell tips For immunoprecipitation experiments from yeast cell extracts, GST- by microtubules. In many eukaryotic systems, proteins asso- mod5p was expressed from the nmt41 promoter to steady-state ciated with microtubules are subsequently targeted to the levels. Tea1p, tea3p-GFP and tea3p-HA were all expressed from their cell cortex, and it is likely that this is achieved by multiple endogenous promoters. A 400–500 mg portion of frozen cell pellets was ground with a mortar and pestle under liquid nitrogen mechanisms (Schuyler and Pellman, 2001; Gundersen, and extracts were prepared in 20 mM NaHEPES pH 7.5, 50 mM 2002; Carvalho et al, 2003; Akhmanova and Hoogenraad, K-acetate, 200 mM NaCl, 1 mM EDTA, 2 mM MgCl and 0.2% Triton 2005; Watanabe et al, 2005). In relation to this, we note that a X-100. A 5–8 mg portion of total protein was used in each key point of our ‘mod5p bypass’ model is that the two immunoprecipitation in a final volume of 500ml. Samples were incubated with 5mg of anti-HA 12CA5, anti-GST or anti-tea1p proposed mechanisms of achieving a tea1p–tea3p interaction antibodies, and 25ml of protein A or protein G sepharose beads as at cell tips are not only complementary but also mutually appropriate, for 2–3 h at 51C. All samples were separated on 8% exclusive. Moreover, the exclusive aspect of their relationship SDS–PAGE and subjected to Western blotting. emerges from the properties of the system as a whole. In wild- type cells, tea1p interacts with tea3p primarily at the cell Physiology and immunofluorescence experiments cortex, because tea3p is sequestered there by mod5p. Immunostaining with anti-tea1p serum and anti-tubulin antibodies was exactly as described (Snaith and Sawin, 2003). Images of entire However, in, and only in, mod5D cells, tea3p is no longer cell volume were captured using a Leica TCS-SP confocal micro- similarly sequestered, and tea1p can therefore interact with scope (Leica Microsystems, Milton Keynes, UK). Fluorescence tea3p in the cytoplasm to form stable complexes that then quantitation of anti-tea1p immunofluorescence (n¼ 130 cells) was travel to the cortex. Importantly, even if the interaction of performed as described (Snaith and Sawin, 2003). For costaining tea3D cells for actin and tea1p, cells were grown in YES and fixed tea1p and tea3p in mod5D cells is less productive than in exactly as described (Sawin and Nurse, 1998), and then digested for wild-type cells (mod5D cells still exhibit some defects in 15 min at 371Cin 25mg/ml zymolyase 20-T. Images were captured polarized growth), both scenarios result in an interaction of using wide-field microscopy on a Nikon TE300 system (Snaith and tea1p and tea3p at the cell cortex, which would then allow Sawin, 2003). Polarity re-establishment experiments were carried tea1p to interact with effector binding partners. In the context out as described (Snaith and Sawin, 2003). In all experiments, 4200 cells were counted for each sample. Treatment with 50mg/ml of ‘systems cell biology’, this may reflect a kind of homeo- MBC was used to disrupt microtubules and 200mM latrunculin B static design principle by which desired outputs can be was used to disrupt actin filaments, during log-phase growth. achieved, even with some degree of fine-tuning, in the face of changing amounts of a critical system component, in this Live cell imaging case mod5p. All images were collected on the Nikon TE300 microscope system. Bright-field imaging of cell growth patterns was performed as described (Glynn et al, 2001). At least 130 cell divisions (i.e., growth Materials and methods patterns in at least 260 daughter cells) were analyzed for each strain. For fluorescence imaging, cells were grown at 251C in EMM General methods and imaged at room temperature. Expression of mutant GFP-mod5p S. pombe methods were as described (Moreno et al, 1991). was induced for 2 days at 321C in EMM plus 150 nM thiamine Molecular biology methods were essentially as described (Sam- before shifting to 251C for 4 h prior to imaging single planes. brook and Russell, 2001). Epitope-tagged mod5þ and tea3þ Z-series of tea3p-GFP still images were deconvolved using Deltavision strains were constructed by PCR-based gene targeting as described Softworx (Applied Precision) and projected into a single plane. (Ba¨hler et al, 1998). Tea3p-GFP, tea3D and tea1D200 strains were Z-series from doubly labeled tea3p-mCh tea1p-GFP or tea3p-mCh gifts from P Nurse, The Rockefeller University, NY. The two-hybrid GFP-atb2p strains were projected into a single plane without further library was a kind gift from T Nakamura, Osaka City University, manipulation. Further details of imaging conditions are given in Japan, and pRSETB-mCherry and pRSETB-tdTomato were gifts of Supplementary data. R Tsien, UCSD, CA. Complete lists of plasmids and strains used in this work appear as Supplementary data. Supplementary data Supplementary data are available at The EMBO Journal Online. Antibodies Anti-GSTantibodies were obtained from Amersham (Little Chalfont, UK); anti-HA 12CA5 antibodies were a gift from I Stancheva, Acknowledgements University of Edinburgh, UK, and anti-TAT1 hybridoma supernatant was a gift from K Gull, University of Oxford, UK (Woods et al, We thank M Masterton for invaluable help in constructing yeast 1989). Affinity-purified tea1p antibodies were used for Western strains expressing mod5p deletion proteins; K Gull, T Nakamura, blotting, and serum was used for immunofluorescence (Snaith and P Nurse, H Ohkura I Stancheva and R Tsien for sharing strains and Sawin, 2003). Anti-tea1p antibodies were purified against a 6-His reagents; A Merdes and I Davis for critical reading of the manu- fusion of tea1p amino acids 554–1147 (tea1p-C) expressed in script; S Martin for helpful discussion of our data; and F Philp bacteria (Mata and Nurse, 1997). Tea1p-C was isolated as inclusion for continuing encouragement. HAS is a Caledonian Research bodies, solubilized in SDS and coupled to Affigel 15 as described Foundation Post-doctoral Research Fellow and KES is a Wellcome (Sawin et al, 1992). Alexafluor-conjugated phalloidin and second- Trust Senior Research Fellow. This work was supported by the ary antibodies were obtained from Molecular Probes (Eugene, OR). Caledonian Research Foundation and the Wellcome Trust. References Adachi Y, Toda T, Niwa O, Yanagida M (1986) Differential expres- Arellano M, Niccoli T, Nurse P (2002) Tea3p is a cell end marker sions of essential and nonessential alpha-tubulin genes in activating polarized growth in Schizosaccharomyces pombe. Curr Schizosaccharomyces pombe. Mol Cell Biol 6: 2168–2178 Biol 12: 751–756 Adams J, Kelso R, Cooley L (2000) The kelch repeat superfamily of Ba¨hler J, Wu JQ, Longtine MS, Shah NG, McKenzie III A, Steever proteins: propellers of cell function. Trends Cell Biol 10: 17–24 AB, Wach A, Philippsen P, Pringle JR (1998) Heterologous Akhmanova A, Hoogenraad CC (2005) Microtubule plus-end-tracking modules for efficient and versatile PCR-based gene targeting in proteins: mechanisms and functions. Curr Opin Cell Biol 17: 47–54 Schizosaccharomyces pombe. Yeast 14: 943–951 3698 The EMBO Journal VOL 24 NO 21 2005 &2005 European Molecular Biology Organization | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al Behrens R, Nurse P (2002) Roles of fission yeast tea1p in the Mitchison JM, Nurse P (1985) Growth in cell length in the fission localization of polarity factors and in organizing the microtubular yeast Schizosaccharomyces pombe. J Cell Sci 75: 357–376 cytoskeleton. J Cell Biol 157: 783–793 Moreno S, Klar A, Nurse P (1991) Molecular analysis of the fission Bretscher A, Edwards K, Fehon RG (2002) ERM proteins and merlin: yeast Schizosaccharomyces pombe. Methods Enzymol 194: integrators at the cell cortex. Nat Rev Mol Cell Biol 3: 586–599 795–823 Browning H, Hackney DD, Nurse P (2003) Targeted movement of Niccoli T, Arellano M, Nurse P (2003) Role of Tea1p, Tea3p and cell end factors in fission yeast. Nat Cell Biol 5: 812–818 Pom1p in the determination of cell ends in Schizosaccharomyces Browning H, Hayles J, Mata J, Aveline L, Nurse P, McIntosh JR pombe. Yeast 20: 1349–1358 (2000) Tea2p is a kinesin-like protein required to generate polar- Prag S, Adams JC (2003) Molecular phylogeny of the kelch-repeat ized growth in fission yeast. J Cell Biol 151: 15–28 superfamily reveals an expansion of BTB/kelch proteins in Brunner D, Nurse P (2000) CLIP170-like tip1p spatially organizes animals. BMC Bioinformatics 4: 42 microtubular dynamics in fission yeast. Cell 102: 695–704 Sambrook J, Russell DW (2001) Molecular Cloning: A Laboratory Busch KE, Hayles J, Nurse P, Brunner D (2004) Tea2p kinesin is Manual. New York: Cold Spring Harbor Laboratory Press involved in spatial microtubule organization by transporting Sawin KE, Mitchison TJ, Wordeman LG (1992) Evidence for tip1p on microtubules. Dev Cell 6: 831–843 kinesin-related proteins in the mitotic apparatus using peptide Carvalho P, Tirnauer JS, Pellman D (2003) Surfing on microtubule antibodies. J Cell Sci 101 (Part 2): 303–313 ends. Trends Cell Biol 13: 229–237 Sawin KE, Nurse P (1998) Regulation of cell polarity by microtu- Chang F, Peter M (2003) Yeasts make their mark. Nat Cell Biol 5: bules in fission yeast. J Cell Biol 142: 457–471 294–299 Sawin KE, Snaith HA (2004) Role of microtubules and tea1p Drubin DG, Nelson WJ (1996) Origins of cell polarity. Cell 84: in establishment and maintenance of fission yeast cell polarity. 335–344 J Cell Sci 117: 689–700 Dyson HJ, Wright PE (2005) Intrinsically unstructured proteins and Schuyler SC, Pellman D (2001) Microtubule ‘plus-end-tracking their functions. Nat Rev Mol Cell Biol 6: 197–208 proteins’: the end is just the beginning. Cell 105: 421–424 Feierbach B, Chang F (2001) Roles of the fission yeast formin for3p Shaner NC, Campbell RE, Steinbach PA, Giepmans BN, Palmer AE, in cell polarity, actin cable formation and symmetric cell division. Tsien RY (2004) Improved monomeric red, orange and yellow Curr Biol 11: 1656–1665 fluorescent proteins derived from Discosoma sp. red fluorescent Feierbach B, Verde F, Chang F (2004) Regulation of a formin protein. Nat Biotechnol 22: 1567–1572 complex by the microtubule plus end protein tea1p. J Cell Biol Small JV, Kaverina I (2003) Microtubules meet substrate adhesions 165: 697–707 to arrange cell polarity. Curr Opin Cell Biol 15: 40–47 Glynn JM, Lustig RJ, Berlin A, Chang F (2001) Role of bud6p and Snaith HA, Sawin KE (2003) Fission yeast mod5p regulates polar- tea1p in the interaction between actin and microtubules for the ized growth through anchoring of tea1p at cell tips. Nature 423: establishment of cell polarity in fission yeast. Curr Biol 11: 836–845 647–651 Gundersen GG (2002) Microtubule capture: IQGAP and CLIP-170 Snaith HA, Sawin KE (2005) Tea for three: control of fission yeast expand the repertoire. Curr Biol 12: R645–R647 polarity. Nat Cell Biol 7: 450–451 Hayles J, Nurse P (2001) A journey into space. Nat Rev Mol Cell Biol Tatebe H, Shimada K, Uzawa S, Morigasaki S, Shiozaki K (2005) 2: 647–656 Wsh3/Tea4 is a novel cell-end factor essential for bipolar dis- Jin H, Amberg DC (2001) Fission yeast Aip3p (spAip3p) is required tribution of Tea1 and protects cell polarity under environmental for an alternative actin-directed polarity program. Mol Biol Cell stress in S. pombe. Curr Biol 15: 1006–1015 12: 1275–1291 Vega LR, Solomon F (1997) Microtubule function in morphological Kim H, Yang P, Catanuto P, Verde F, Lai H, Du H, Chang F, Marcus S differentiation: growth zones and growth cones. Cell 89: 825–828 (2003) The kelch repeat protein, Tea1, is a potential substrate Verde F (2001) Cell polarity: a tale of two Ts. Curr Biol 11: target of the p21-activated kinase, Shk1, in the fission yeast, R600–R602 Schizosaccharomyces pombe. J Biol Chem 278: 30074–30082 Verde F, Mata J, Nurse P (1995) Fission yeast cell morphogenesis: Li X, Zhang D, Hannink M, Beamer LJ (2004) Crystal structure identification of new genes and analysis of their role during the of the Kelch domain of human Keap1. J Biol Chem 279: cell cycle. J Cell Biol 131: 1529–1538 54750–54758 Watanabe T, Noritake J, Kaibuchi K (2005) Regulation of micro- Martin SG, Chang F (2003) Cell polarity: a new mod(e) of anchor- tubules in cell migration. Trends Cell Biol 15: 76–83 ing. Curr Biol 13: R711–R713 Woods A, Sherwin T, Sasse R, MacRae TH, Baines AJ, Gull K (1989) Martin SG, McDonald WH, Yates III JR, Chang F (2005) Tea4p links Definition of individual components within the cytoskeleton microtubule plus ends with the formin for3p in the establishment of Trypanosoma brucei by a library of monoclonal antibodies. of cell polarity. Dev Cell 8: 479–491 J Cell Sci 93: 491–500 Mata J, Nurse P (1997) Tea1p and the microtubular cytoskeleton are Yarm F, Sagot I, Pellman D (2001) The social life of actin important for generating global spatial order within the fission and microtubules: interaction versus cooperation. Curr Opin yeast cell. Cell 89: 939–949 Microbiol 4: 696–702 &2005 European Molecular Biology Organization The EMBO Journal VOL 24 NO 21 2005 3699 | | http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The EMBO Journal Springer Journals

Multistep and multimode cortical anchoring of tea1p at cell tips in fission yeast

Download PDF
10 pages

Loading next page...
 
/lp/springer-journals/multistep-and-multimode-cortical-anchoring-of-tea1p-at-cell-tips-in-Atvt0naRBW

References (52)

Publisher
Springer Journals
Copyright
Copyright © European Molecular Biology Organization 2005
ISSN
0261-4189
eISSN
1460-2075
DOI
10.1038/sj.emboj.7600838
Publisher site
See Article on Publisher Site

Abstract

The EMBO Journal (2005) 24, 3690–3699 & 2005 European Molecular Biology Organization All Rights Reserved 0261-4189/05 | | THE THE www.embojournal.org EMB EMB EMBO O O JO JOU URN R NAL AL Multistep and multimode cortical anchoring of tea1p at cell tips in fission yeast skeleton (Hayles and Nurse, 2001; Chang and Peter, 2003). Hilary A Snaith, Itaru Samejima After a variety of stresses, tea1D cells become bent or and Kenneth E Sawin* branched (Verde et al, 1995; Mata and Nurse, 1997; Niccoli Wellcome Trust Centre for Cell Biology, University of Edinburgh, et al, 2003; Sawin and Snaith, 2004), in part because micro- Edinburgh, UK tubules are unable to specify positional information to the cell cortex (Sawin and Nurse, 1998; Feierbach et al, 2004; The fission yeast cell-polarity regulator tea1p is targeted Sawin and Snaith, 2004). In addition, during steady-state to cell tips by association with growing microtubule ends. growth, tea1D mutants exhibit a mostly monopolar growth Tea1p is subsequently anchored at the cell cortex at pattern, unlike wild-type cells, which exhibit monopolar cell tips via an unknown mechanism that requires both growth early in the cell cycle and bipolar growth later in the tea1p carboxy-terminus and the membrane protein the cell cycle (Mitchison and Nurse, 1985; Mata and Nurse, mod5p. Here, we show that a tea1p-related protein, 1997; Glynn et al, 2001). tea1D mutants are also defective tea3p, binds independently to both mod5p and tea1p, in the cortical localization of other polarity factors, including and that tea1p and mod5p can also interact directly, the actin-binding protein bud6/aip3 (Glynn et al, 2001; independent of tea3p. Despite their related structures, Jin and Amberg, 2001), the formin for3p (Feierbach and different regions of tea1p and tea3p are required for their Chang, 2001; Feierbach et al, 2004), the SH3-domain protein respective interactions with an essential central region of tea4p/wsh3p (Martin et al, 2005; Tatebe et al, 2005) and the mod5p. We demonstrate that tea3p is required for proper tea1p-related protein tea3p (Arellano et al, 2002). cortical localization of tea1p, specifically at nongrowing Tea1p is a 1147-amino-acid protein. Its N-terminus con- cell tips, and that tea1p and mod5p are independently tains six kelch repeats, which are found in many proteins, required for tea3p localization. Further, we find that tea3p including actin-binding proteins, and are likely involved in fused to GFP or mCherry is cotransported with tea1p by protein–protein interactions (Adams et al, 2000; Prag and microtubules to cell tips, but this occurs only in the Adams, 2003; Li et al, 2004). The C-terminal regions of tea1p absence of mod5p. These results suggest that independent are predicted to be largely alpha-helical coiled coil. Tea1p also protein–protein interactions among tea1p, tea3p and displays limited functional and sequence similarity to ezrin, mod5p collectively contribute to tea1p anchoring at cell a member of the ERM (ezrin–radixin–moesin) family, which tips via a multistep and multimode mechanism. has been shown to link the plasma membrane with the The EMBO Journal (2005) 24, 3690–3699. doi:10.1038/ underlying actin cytoskeleton in animal cells (Vega and sj.emboj.7600838; Published online 13 October 2005 Solomon, 1997; Bretscher et al, 2002). Subject Categories: cell & tissue architecture In vivo, tea1p displays a dynamic pattern of localization, Keywords: cell polarity; kelch repeat; mCherry; S. pombe; reflecting its function in cell polarity. Tea1p is transported to tdTomato cell tips on the growing plus ends of microtubules (Behrens and Nurse, 2002; Snaith and Sawin, 2003; Feierbach et al, 2004), dependent on the kip2-like kinesin tea2p and the CLIP- 170 homologue tip1p (Browning et al, 2000, 2003; Brunner Introduction and Nurse, 2000; Busch et al, 2004). Once at tips, tea1p is unloaded from microtubules and retained at the cell cortex. Communication between microtubules and the actin cyto- This retention is defective in mutants lacking either the skeleton is a common feature of eukaryotic cell polarity membrane protein mod5p or the C-terminal 200 amino (Drubin and Nelson, 1996; Yarm et al, 2001; Chang and acids of tea1p (tea1D200; Behrens and Nurse, 2002; Martin Peter, 2003; Small and Kaverina, 2003). In many cases, and Chang, 2003; Snaith and Sawin, 2003; also see below). proteins that are transported on the plus ends of micro- Cortically associated tea1p then contributes to the organiza- tubules, such as EB1 and CLIP-170, forge interactions with tion of actin filaments at cell tips via interactions with proteins proteins localized at sites on the plasma membrane, facilitat- such as tea4p/wsh3p and bud6p/aip3p, thereby ensuring ing this communication (Schuyler and Pellman, 2001; proper bipolar growth (Glynn et al, 2001; Verde, 2001; Gundersen, 2002; Carvalho et al, 2003; Akhmanova and Martin et al, 2005; Snaith and Sawin, 2005; Tatebe et al, 2005). Hoogenraad, 2005; Watanabe et al, 2005). Currently, there are three major outstanding issues in In the fission yeast Schizosaccharomyces pombe, tea1p is relation to how tea1p functions in microtubule-mediated a key mediator between microtubules and the actin cyto- cell polarity in fission yeast. At the beginning of the tea1p ‘pathway’ is the question of how tea1p is associated with and *Correspondence: Wellcome Trust Centre for Cell Biology, Swann Building, School of Biological Sciences, University of Edinburgh, transported on microtubule plus ends, and at the end of the Mayfield Road, Edinburgh EH9 3JR, UK. Tel.: þ 44 131 650 7064; pathway is the question of how tea1p interacts with the actin Fax: þ 44 131 650 7360; E-mail: [email protected] cytoskeleton. Here, we address a question at the center of the tea1p pathway, namely, how does tea1p become anchored at Received: 21 March 2005; accepted: 15 September 2005; published online: 13 October 2005 the cortex at cell tips? 3690 The EMBO Journal VOL 24 NO 21 2005 &2005 European Molecular Biology Organization | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al An important factor in tea1p anchoring is the protein A α-GST IP WCE mod5p, which was first identified in a screen for mutations + Δ + + + Δ + + GST-mod5 in S. pombe that affect cell shape (Snaith and Sawin, 2003). tea3-GFP + + Δ + + + Δ + In mod5D cells, tea1p is transported normally to cell tips but + + + Δ + + + Δ tea1 fails to accumulate there (Snaith and Sawin, 2003). Mod5p, GST-mod5p which is unusually rich in serine, threonine and proline residues and thus may be an intrinsically unstructured pro- tea3p-GFP tein (Dyson and Wright, 2005), is localized to membranes at tea1p cell tips, via a C-terminal prenylation sequence. Interestingly, in tea1D cells, mod5p is no longer restricted to cell tips but 13 2 4 5 6 7 8 rather spreads out around the entire cell cortex. This has led α-HA IP WCE to the suggestion of a positive feedback loop linking mod5p B mod5 + + Δ + + Δ and tea1p localization, which might ensure high-fidelity tea3-HA – + + – + + polarized growth in the context of dynamic tea1p delivery tea1 + + + + + + to cell tips (Snaith and Sawin, 2003). However, the molecular mechanisms underlying tea1p localization at the cell cortex tea3p-HA remain largely unclear. For example, although there is genetic evidence for an interaction between mod5p and tea1p, a clear tea1p biochemical interaction has not yet been shown. In particu- 1324 5 6 lar, live-cell analyses of GFP-tagged mod5p and tea1p have demonstrated that the two proteins do not completely co- C kelch repeats Coiled coil localize at cell tips, and that tea1p at the cortex is highly tea1p 1 1147 78 370 609 1106 dynamic (Snaith and Sawin, 2003). Thus, even if tea1p and mod5p Membrane association/ binding tea3p binding mod5p do interact physically, whether directly or indirectly, Coiled coil other proteins may also be involved in regulating their kelch repeats tea3p 1 1125 localization. 179 371 599 966 In this work, we have sought to further understand the mod5p tea1p binding binding mechanisms regulating tea1p localization at cell tips by Prenylation identifying protein–protein interactions involving mod5p. signal From a two-hybrid screen, we identified the tea1p-related mod5p 1 522 protein tea3p as a mod5-interacting protein. In these experi- tea1p/tea3p binding ments, we also found that mod5p interacts with tea1p and that tea1p interacts with tea3p. In vivo, each of the pairwise WCE GST-IP tea1p-IP interactions occurring between any two of these three pro- tea1 + Δ200 + Δ200 + Δ200 teins can take place in the absence of the third protein, and in a series of localization-dependency studies, we found that GST-mod5 each of these proteins contributes to the proper localization tea1p tea1Δ200p and function of the other two, but in distinct ways. These results suggest that a complex network of interactions among tea3p-HA tea1p, mod5p and tea3p is involved in regulating the locali- 13 2 4 5 6 zation of these three proteins, ultimately leading to correctly Figure 1 Tea1p, mod5p and tea3p form independent complexes anchored tea1p at cell tips. in vivo.(A) GST-mod5p was immunoprecipitated from soluble protein extracts of wild-type cells expressing GST-mod5p and tea3p-GFP (lanes 1 and 5), mod5D cells expressing tea3p-GFP (lanes 2 and 6), Results tea3D cells expressing GST-mod5p (lanes 3 and 7) and tea1D cells expressing GST-mod5p and tea3p-GFP (lanes 4 and 8). The resulting Pairwise interactions between tea1p, tea3p and mod5p immunocomplexes were analyzed for GST-mod5p, tea3p-GFP and To better understand the mechanisms underlying tea1p loca- tea1p. Whole-cell extract (WCE) fractions are shown in lanes 1–4 lization. we performed a two-hybrid screen using essentially and immunoprecipitates in lanes 5–8. Immunoprecipitates were full-length mod5p as bait (see Supplementary data). Out of 11 loaded 30 relative to WCE sample. (B) Tea3p-HA was immuno- precipitated from soluble protein extracts of wild-type cells (lanes 1 plasmids isolated, 10 contained fragments of the tea3 gene, and 4), wild-type cells expressing tea3p-HA (lanes 2 and 5) and previously implicated in cell polarity (Arellano et al, 2002). mod5D cells expressing tea3p-HA (lanes 3 and 6). The resulting Interestingly, tea3p is structurally related to tea1p. Both immunocomplexes were analyzed for tea3p-HA and tea1p. WCE proteins contain the protein–protein interaction kelch domain fractions are shown in lanes 1–3 and immunoprecipitates in lanes 4–6. Immunoprecipitates were loaded 20 relative to WCE. in their N-termini (Adams et al, 2000; Prag and Adams, 2003; (C) Schematic diagram summarizing interactions between tea1p, tea3p Li et al, 2004) and long regions of predicted coiled coil in and mod5p. (D) GST-mod5p (lanes 3 and 4) or tea1p (lanes 5 and 6) their C-termini. We confirmed the interaction between mod5p was immunoprecipitated from soluble protein extracts of either and tea3p in GST pulldown experiments and co-immuno- wild-type (lanes 1, 3 and 5) or tea1D200 (lanes 2, 4 and 6) cells expressing GST-mod5p and tea3p-HA. The resulting immunocom- precipitation assays, using extracts from fission yeast plexes were analyzed for GST-mod5p, tea1p (or tea1D200p) and simultaneously expressing GST-mod5p and tea3p-GFP tea3p-HA. WCE fractions are shown in lanes 1 and 2 and immuno- (Supplementary Figure 1A and Figure 1A). precipitates in lanes 3–6. Immunoprecipitates were loaded 30 Arellano et al (2002) reported that tea3p bound to tea1p relative to WCE. in the yeast two-hybrid system, but they did not verify this &2005 European Molecular Biology Organization The EMBO Journal VOL 24 NO 21 2005 3691 | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al biochemically. We found that immunoprecipitation of tea3p- Full-length or mutant GFP-mod5p expression in: HA from fission yeast cell extracts co-precipitated tea1p mod5Δ background mod5 + bkgd (Figure 1B). In addition, we found that tea1p was co- tea1p/tubulin GFP signal tea1p/tubulin precipitated in immunoprecipitates from cells expressing A EI GST-mod5p (Figure 1A). Thus, tea1p, tea3p and mod5p all associate with each other in vivo. Based on these results, we wanted to examine whether tea3p might act as a molecular ‘bridge’ between tea1p and mod5p. We therefore used deletion strains to test whether GFP-mod5p full-length each of the pairwise protein–protein interactions observed in immunoprecipitation experiments could occur in the absence B F J of the third protein. As expected from two-hybrid results, tea1p interacted with tea3p in the absence of mod5p, and mod5p interacted with tea3p in the absence of tea1p (Figure 1A and B). Interestingly, however, we also found that tea1p interacted with mod5p in the absence of tea3p (Figure 1A). GFP-mod5Δ256 – 305p Thus, each of the pairwise interactions among tea1p, tea3p and mod5p is independent of the third protein. C G K Through a combination of immunoprecipitation experi- ments, two-hybrid analysis and in vitro binding studies, we mapped the regions of tea1p, tea3p and mod5p involved in binding to each other (Supplementary Figures 1B–F, 2A and B). The results are summarized in Figure 1C. Four important GFP-mod5Δ156 – 205p points emerged from these experiments. First, all of the observed interactions are likely to be direct. Second, a central D H L region of mod5p (amino acids 156–205) is required for binding to both tea1p and tea3p. Third, even though tea1p and tea3p are structurally related, binding of tea1p to mod5p is mediated by the N-terminus of tea1p (amino acids 1–352), while binding of tea3p to mod5p is mediated by a central GFP-mod5Δ206 – 255p coiled-coil region of tea3p (amino acids 739–785). Fourth, binding of tea1p to tea3p is mediated by the C-termini of Figure 2 Mod5p amino acids 156–255 are essential for the localiza- both proteins (amino acids 948–1147 of tea1p and amino tion of tea1p and mod5p. The localization of tea1p (green) and microtubules (red) in (A–D) mod5D cells and (I–L) wild-type cells, acids 901–1125 of tea3p). These last two points are particu- expressing different mutant versions of GFP-mod5p. (E–H) The larly salient because Behrens and Nurse (2002) have shown localization of GFP-mod5p (and mutant versions) in mod5D cells. that deletion of the tea1p C-terminus (tea1D200) prevents (A, E, I) Wild-type GFP-mod5p; (B, F, J) GFP-mod5D256–305p; anchoring of tea1p to the cell cortex. In this context, our (C, G, K) GFP-mod5D156–205p; (D, H, L) GFP-mod5D206–255p. The scale bar represents 5mm. mapping data suggest that the anchoring defect of tea1D200 mutants correlates not with a failure to bind mod5p but rather with a failure to bind tea3p and/or other proteins (see Discussion). The central region of mod5p is required for function We next tested whether tea1p, tea3p and mod5p all coexist We next wanted to determine to what extent the protein– in a single protein complex in vivo. This was not entirely protein interactions identified might mediate the localization straightforward to determine, because physical analysis of and/or function of mod5p, tea3p and tea1p, in order to tea1p shows it to be present in complexes covering a wide understand how these interactions contribute to the proper range of molecular sizes (Feierbach et al, 2004). Moreover, cortical anchoring of tea1p at cell tips. because each of the three proteins concerned can interact The only recognizable amino-acid sequence motif in independently with the other two, immunoprecipitation of mod5p is a C-terminal prenylation signal that is essential for any one protein will co-precipitate both of the other two, even mod5p function and localization (Snaith and Sawin, 2003). To if a three-way complex does not exist. We therefore examined identify other functionally significant regions, we fused GFP to whether a three-way complex could be identified when one a series of 50-amino-acid internal deletions spanning the of the pairwise interactions was disrupted. Interestingly, we mod5p open reading frame (i.e., those used in mapping found that tea3p could not be co-immunoprecipitated with studies; Supplementary Figure 2A) and expressed the inter- tea1D200p, even though both proteins could bind to mod5p nal-deletion mutant proteins individually in mod5D cells. In a in the same cell extract (Figure 1D). This suggests that a quantitative polarity maintenance assay, nearly all of the three-way complex of tea1p, tea3p and mod5p may be deletions behaved like wild-type mod5p; only cells expressing present only transiently in vivo, if at all. However, we also either mod5D156–205p or mod5D206–255p, which fail to bind found that the absence of a three-way complex in vivo is tea1p or tea3p (Supplementary Figure 2A and B), were unable unlikely to be a result of tea1p and tea3p competing for to maintain polarity (Supplementary Figure 3A). potentially overlapping binding sites on mod5p, as a three- We also studied the localization of tea1p and the GFP- way complex could be demonstrated artificially, in a yeast mod5p internal-deletion proteins themselves in each internal- ‘bridging two-hybrid’ assay (Supplementary Figure 2C). deletion mutant strain. In nearly all cases, both tea1p and the 3692 The EMBO Journal VOL 24 NO 21 2005 &2005 European Molecular Biology Organization | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al mutant versions of mod5p were localized as in wild-type cells tea3p-GFP (Figure 2A, B, E and F; additional data not shown). How- A B C ever, in cells expressing either mutant mod5D156–205p or mod5D206–255p, tea1p localization resembled that of mod5D cells, with tea1p present on microtubule plus ends but failing to accumulate at the cell cortex (Figure 2C and D and Supplementary Figure 4E). These same two mutants were also defective in mod5 localization (Figure 2G and H), Wild type tea1ΔΔ tea1 200 with the mutant GFP-mod5p spread out around the plasma D E F membrane in the same manner as full-length wild-type mod5p in tea1D cells (Snaith and Sawin, 2003). Tea1p localization was not altered from wild type when any of the internal-deletion proteins was expressed in mod5 rather than mod5D backgrounds, even though the localization of GFP-tagged mod5D156–205p and mod5D206–255p remained mod5ΔΔ mod5 tea1ΔΔ mod5 206 – 255 aberrant in these experiments (Figure 2I–L; additional data Figure 3 Localization dependencies of tea3p-GFP. Localization of not shown). These results suggest that, in addition to the tea3p-GFP in (A) wild-type, (B) tea1D,(C) tea1D200,(D) mod5D, C-terminal prenylation site, the central region of mod5p, (E) mod5D tea1D and (F) mod5D206–255 cells. The scale bar represents 5mm. which interacts with both tea1p and tea3p, contains the major determinants for both mod5p localization and function. We previously showed that restriction of mod5p to cell tips is dependent on tea1p and also partially on tea3p notion that tea1p can promote the localization of tea3p to cell (Supplementary Figure 3B and C; Snaith and Sawin, 2003). tips, while mod5p acts to stably integrate tea3p at the cell In light of our results showing that mod5p can bind to cortex, to the extent that the mislocalization of mod5p (i.e., tea1D200p, we examined the localization of GFP-mod5p in in tea1D mutants) will recruit tea3p to ectopic cortical sites. tea1D200 cells. GFP-mod5p was spread out around the We also found that in cells expressing mod5D156–205p membrane (Supplementary Figure 3D), suggesting that or mod5D206–255p, tea3p-GFP was delocalized in a pattern restriction of mod5p to cell tips requires not only the similar to that present in mod5D cells (Figure 3F; data not tea1p–mod5p interaction but also the stable binding of shown). In conclusion, the overall tip localization of tea3p is tea1p at the cell cortex. dependent both on binding to mod5p and on the carboxy- terminus of tea1p. Tea1p and mod5p are independently required for different aspects of tea3p localization We next sought to investigate the roles played by tea1p and Microtubule-based transport of tea3p to cell tips mod5p in the localization of tea3p. In wild-type cells, the requires both tea1p and the absence of mod5p majority of tea3p-GFP was confined to the cell tip region, Because tea1p binds to tea3p and facilitates its localization to with a few cytoplasmic dots also present, confirming pre- cell tips, and tea1p is transported to cell tips by association vious results (Figure 3A; Arellano et al, 2002). In contrast, in with microtubule plus ends (Behrens and Nurse, 2002; Snaith tea1D cells, tea3p-GFP was still partially enriched at cell tips and Sawin, 2003; Feierbach et al, 2004), we investigated but also displayed a punctate staining evenly spread around whether tea3p is similarly transported to cell tips along the plasma membrane (Figure 3B; Arellano et al, 2002), microtubules. In wild-type cells, tea3p-GFP displayed little reminiscent of GFP-mod5p localization in tea1D mutants or no apparent movement toward cell tips (Supplementary (Snaith and Sawin, 2003). In addition, we observed an Movie 1). Previously, however, we demonstrated that levels increase in the cytoplasmic tea3p-GFP signal (Figure 3B). In of tea1p-GFP transport to cell tips are elevated in mod5D cells tea1D200 cells, the same mislocalization of tea3p-GFP was relative to wild-type cells, and we argued that this might be seen (Figure 3C), consistent with the carboxy-terminal region due to the fact that an inability of tea1p to anchor at the of tea1p being required for binding to tea3p (Supplementary cortex could result in a higher free cytoplasmic pool of tea1p Figure 1F). able to associate with microtubules (Snaith and Sawin, 2003). Because mod5p is required for proper tea1p localization, We therefore followed tea3p-GFP in mod5D cells. and tea1p is required for tea3p localization, we suspected that Here, we observed significant motion of cytoplasmic tea3p- mod5p would also be necessary for the correct localization GFP particles. Most of the fainter particles of tea3p-GFP of tea3p. Interestingly, however, the manner of tea3p-GFP showed rapid movements with no clear directionality. mislocalization in mod5D mutants was different from that in However, many of the bright particles of tea3p-GFP exhibited tea1D mutants. In mod5D cells, the tea3p-GFP signal became linear movements, both toward and away from cell tips highly punctate, with a few bright dots of tea3p-GFP visible (Figure 4A and Supplementary Movie 2). To test whether near cell tips and also throughout the cell, suggesting that this movement was dependent on the actin or microtubule mod5p is required for a stable membrane localization cytoskeletons, we treated cells with latrunculin B or methyl of tea3p (Figure 3D). Consistent with this, we found that 2-benzimidazolecarbamate (carbendazim; MBC). Treatment tea3p-GFP localization in mod5D tea1D double mutants of cells with latrunculin B for up to an hour had no significant resembled tea3p-GFP in mod5D single mutants, although effect on the movement of tea3p-GFP (data not shown). By tea3p-GFP localization to cell tips was more compromised contrast, within 5 min of MBC treatment, the movement of than in single mutants (Figure 3E). These results support the bright tea3p-GFP particles was reduced, and after 20 min it &2005 European Molecular Biology Organization The EMBO Journal VOL 24 NO 21 2005 3693 | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al mod5Δ mod5Δ Wild type tea3p-GFP GFP-atb2 /tea3p-mCh GFP-atb2 /tea3p-mCh no MBC plus MBC A B C Wild type tea1p-GFPt /tea3p-mCh Wild type 88 – 12 – mod5Δ 57 37 6 – tea3Δ 15 77 5 – tea1Δ 12 76 – 8 Figure 5 Mod5p and tea3p have distinct but overlapping functions. (A) Wild-type, mod5D, tea3D, tea1D, mod5D tea3D, mod5D tea1D, mod5ΔΔ mod5 tea1Δ mod5Δ tea1p-GFP/tea3p-mCh tea3p-GFP tea3D tea1D and tea1D200 cells were depolarized by growth to tea1p-GFP/ tea3p-mCh stationary phase and returned to fresh medium in the absence (gray E F G bars) or presence (black bars) of 50mg/ml MBC for 3 h at 321C. The percentage of branched cells in each sample was counted, n¼ 200. (B) Percentage of daughter cell pairs displaying illustrated initial growth patterns after septation in wild-type (n¼ 133), mod5D (n¼ 228), tea3D (n¼ 165) and tea1D cells (n¼ 194). Arrows indicate direction of growth. conjunction with GFP-tubulin (GFP-atb2p; Adachi et al, 1986). In wild-type cells, tea3p-mCh accumulated at cell tips, consistent with the localization of tea3p-GFP, although in further experiments we found that fusion of tea3p with mCherry may in fact compromise some more subtle aspects of tea3p behavior as compared to fusion with GFP (data not shown). When we imaged tea3p-mCh together with GFP- tubulin in mod5D mutants, tea3p-mCh particles were readily detected on the ends of growing microtubules (Figure 4B). In wild-type cells, however, tea3p-mCh particles failed to colocalize with microtubule ends (Figure 4C). Figure 4 Microtubule-dependent movement of tea3p in mod5D. We next used tea3p-mCh to test whether tea3p was co- (A) Time-lapse movie frames of tea3p-GFP in mod5D cells at 15 s intervals. Red and turquoise arrowheads mark traveling particles of transported with tea1p to cell tips. In wild-type cells, some tea3p-GFP. (B) Time-lapse movie frames of GFP-atb2p (green) and cortical particles of tea3p-mCh colocalized with tea1p-GFP, tea3p-mCh (red) in mod5D cells at 15 s intervals. Gray dashed line while others remained distinct from tea1p-GFP (Figure 4D), indicates the starting position of the traveling tea3p-mCh particle. (C) Localization of GFP-atb2p (green) with tea3p-mCh (red) in wild- suggesting that there may be distinct populations of each type cells. (D) Localization of tea1p-GFP (green) with tea3p-mCh protein, with an overlapping subset. In mod5D cells, particles (red) in wild-type cells. (E, F) Time-lapse movie frames of tea1p- of tea3p-mCh colocalized with tea1p-GFP as they translocated GFP (green) and tea3-mCh (red) in mod5D cells at 15 intervals. both toward and away from the cell tips (Figure 4E and F). Traveling particles of colocalized tea1p-GFP and tea3p-mCh are (As described in Supplementary data, relatively long expo- indicated by white arrowheads and static particles are indicated by white arrows. (G) Time-lapse movie frames of tea3p-GFP in sure times were required during time-lapse acquisition, re- mod5 tea1D cells at 15 s intervals. Red dotted lines indicate reduced sulting in occasional misalignment of the moving tea3p-mCh movement of tea3p-GFP. The scale bar represents 5mm. and tea1p-GFP signals.) Thus, tea3p and tea1p can move together. Finally, we tested whether the movement of tea3p-GFP was almost abolished (data not shown), suggesting that tea3- seen in mod5D cells was dependent on tea1p. In mod5D GFP was moving on microtubules. tea1D double mutant cells, the linear movement of bright To confirm this, we developed a series of tagging plasmids particles of tea3p-GFP was almost completely abolished derived from novel variants of the red fluorescent protein (Figure 4G and Supplementary Movie 3). Collectively, these dsRed (Shaner et al, 2004; see Supplementary data) and used results suggest that tea3p does not significantly associate these to construct strains carrying tea3p tagged with the with tea1p or microtubules in the cytoplasm in wild-type variant mCherry (tea3p-mCh), in order to image tea3p in cells. However, in mod5D cells, where tea3p is no longer 3694 The EMBO Journal VOL 24 NO 21 2005 &2005 European Molecular Biology Organization | | % branched cells Wild type mod5Δ tea3Δ tea1 mod5Δ tea3Δ mod5Δ tea1 tea3Δ tea1Δ tea1Δ200 Tea1p, mod5p and tea3p in fission yeast HA Snaith et al sequestered at cells tips, tea3p can associate with tea1p in the delivery of tea1p to cell tips (Snaith and Sawin, 2003). We cytoplasm and as a result is cotransported with tea1p along found that tea3D cells, like mod5D cells, formed branches microtubules to cell tips. only in the presence of MBC (Figure 5A). If tea3p and mod5p both acted in a strictly linear genetic pathway to regulate cell Mod5p and tea3p have distinct but overlapping polarity, we might expect that the phenotype of the mod5D functions in cell polarity and tea1p anchoring tea3D double mutant would resemble the single mutants. Both mod5D and tea3D mutants have been shown to have However, a significant fraction of mod5D tea3D mutants growth-polarity defects, but thus far each mutant has been (20%) formed branched cells even without microtubule dis- examined in a distinct type of polarity assay (Arellano et al, ruption, much more than occurs in either single mutant 2002; Snaith and Sawin, 2003). To determine whether mod5p (Figure 5A). This suggests that although mod5p and tea3p and tea3p act in a single common pathway, we tested several both interact and contribute to growth polarity in fission mutants side by side in both assays. yeast, each protein may make a distinct contribution to cell In return-to-growth experiments involving cells previously polarity and tea1p function. grown to stationary phase, wild-type cells re-establish polar- In these experiments, we also found that tea1D200 cells ity axes at pre-existing cell ends, and tea1D mutants form formed branches both in the presence and absence of MBC branched cells (Browning et al, 2000; Snaith and Sawin, (Figure 5A) and thus resemble tea1D rather than mod5D 2003). By contrast, mod5D mutants re-establish polarity or tea3D mutants with regard to polarity defects. Because normally when microtubules are intact but form branches tea1D200p can still bind to mod5p (Supplementary Figure when microtubules are disrupted with MBC, which impairs 1E), this suggests that the essential function of the C-terminus A B C – MBC Wild type tea3 Δ orb2-34 D E F + MBC Wild type tea3 Δ orb2-34 tea3Δ G H Wild type 01 5 0 15 20 01 5 0 15 20 Time in MBC (min) Time in MBC (min) neither end 1 end 2 ends neither end 1 end 2 ends I J K – MBC L M N + MBC Figure 6 Tea1p is preferentially lost from the nongrowing tip in tea3D cells. (A–H) Treatment of (A, D) wild-type, (B, E) tea3D and (C, F) orb2- 34 cells with MBC for (A–C) 0 min or (D–F) 5 min. Cells are stained for tea1p (green) and microtubules (red). Time course of tea1p loss from cell tips in (G) wild-type and (H) tea3D cells, showing number of cells with detectable levels of tea1p at one cell tip, two tips or neither tip, for each strain. n¼ 200 for each strain. (I–N) Treatment of tea3D cells with MBC for (I–K) 0 min or (L–N) 5 min. Cells are stained with anti-tea1p antibodies (I, L) and Alexa-labeled phalloidin (J, M). Merged imaged are shown with tea1p in green and phalloidin in red (K, N). The scale bars represent 5mm. &2005 European Molecular Biology Organization The EMBO Journal VOL 24 NO 21 2005 3695 | | Number of cells Number of cells Tea1p, mod5p and tea3p in fission yeast HA Snaith et al of tea1p in anchoring tea1p may be required subsequent to symmetrically from both cell tips (Figure 6A and D), tea3D the tea1p–mod5p interaction. In addition, because tea1D200p cells rapidly lost tea1p from one tip, such that after 5 min of fails to bind tea3p but the tea1D200 phenotype is more MBC treatment, nearly all tea3D cells displayed monopolar extreme than the tea3D phenotype, this indicates that the tea1p localization (Figure 6B and E). To determine whether tea1p carboxy-terminus is likely to interact functionally not this asymmetric loss of tea1p was simply due to the mono- only with tea3p but also with other proteins as well (see polar growth pattern of tea3D cells, we assayed tea1p loca- Discussion). lization in the pak1/shk1/orb2 mutant orb2-34, which also The second assay involved time-lapse observations of exhibits strongly monopolar growth (Verde et al, 1995; Kim growth patterns in the different mutants. Whereas tea3D et al, 2003). In contrast to tea3D cells, orb2-34 cells lost tea1p cells mainly exhibit monopolar growth, in a pattern similar symmetrically after treatment with MBC (Figure 6C and F). to tea1D mutants (Glynn et al, 2001; Arellano et al, 2002), we We also examined actin and tea1p localization in MBC- found that the majority of mod5D cells display wild-type treated tea3D cells to determine whether tea1p was lost growth patterns, with less than 50% of cells showing the from the growing tip, the nongrowing tip or randomly from aberrant growth patterns characteristic of tea1D and tea3D either tip (Figure 6I–N). In nearly all tea3D cells, tea1p was mutants (Figure 5B). This result strongly supports a view lost from the nongrowing, actin-poor tip; the small number of of overlapping but distinct functions for mod5p and tea3p. cases in which loss of tea1p was symmetrical (7%, n¼ 100) Moreover, since the monopolar growth defects in mod5D corresponded to those few tea3D cells that had initiated mutants are much less severe than in tea1D mutants, these bipolar growth. These results indicate that tea3p is required data also suggest that at least some ‘bipolar growth’ functions for the proper cortical anchoring of tea1p, specifically at of tea1p can bypass the need for mod5p at cell tips (see nongrowing cell tips. Discussion). While we have shown that mod5p is required for anchor- Discussion ing of tea1p at cell tips (Snaith and Sawin, 2003), Arellano et al (2002) reported that tea3p does not play a role in tea1p Multistep and multimode tea1p anchoring localization. Our studies with these proteins prompted us to We have demonstrated multiple interactions among tea1p, re-examine the localization of tea1p in tea3D mutants. We tea3p and mod5p and examined their roles in cell polarity found that although tea1p is present at both tips of tea3D and in regulating protein localization at cell tips, especially cells, tea3D cells have only 76% of the tea1p levels seen at that of tea1p. Because interactions between any two of these cell tips in wild-type cells, while tea1p localization in mod5D three proteins can occur in vivo in the absence of the third tea3D double mutants resembles tea1p localization in mod5D protein, while mutation and/or deletion of any one protein cells (Supplementary Figure 4). Because this reduction of tip- characteristically alters the localization of each of the other localized tea1p in tea3D mutants could reflect a role for tea3p two, it might appear difficult to place these interactions at in anchoring tea1p at cell tips, we examined tea1p levels at specific points in a pathway regulating tea1p anchoring. cell tips in wild-type cells and tea3D mutants after micro- However, by taking into account our additional findings— tubule disruption by MBC, where new tea1p delivery to cell for example, the asymmetric cell-tip loss of tea1p in tea3D tips is prevented. Under these conditions, tea1p is rapidly lost mutants, the cotransport of tea3p with tea1p only in the from both cell tips in mod5D mutants (HA Snaith, unpub- absence of mod5p and the behavior of tea1D200p—we have lished data). Whereas wild-type cells gradually lost tea1p formulated a multistep and multimode model for the efficient Nongrowing cell tip tea1p tea3p C tea3p mod5p anchors tea1p and tea3p tea3p at cortex 'bypass' mod5p mod5p tea1p tea1p interacts (mod5Δ) with tea3p tea3p A B tea1p tea1p tea1p tea1p delivered by mod5p MTs to cortex tea1p interacts tea1p localization at with mod5p at both cell tips both cell tips tea1p tea1p interacts with X (= tea4p?) Growing cell tip Figure 7 Multistep cortical retention of tea1p. A model of protein–protein interactions regulating tea1p localization at cell tips. (A) Tea1p arrives at the cell cortex on microtubule plus ends. (B) The tea1p N-terminus interacts with mod5p at both cell tips. (C) At nongrowing cell tips, the tea1p C-terminus interacts with tea3p, which is itself anchored at the cortex (D) via interaction with mod5p. (E) At growing cell tips, tea1p interacts with unknown factors (X) that provide a function similar to that of tea3p at nongrowing tips. (F) A partially functional interaction between tea1p and tea3p can occur without mod5p, but tea1p and tea3p will be poorly anchored. (G) Together, these interactions lead to proper tea1p anchoring at both cell tips. See text for further details. 3696 The EMBO Journal VOL 24 NO 21 2005 &2005 European Molecular Biology Organization | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al anchoring of tea1p at cell tips (Figure 7). The model is interact with tea3p, we have placed a tea1p–tea3p interaction ‘multistep’ because it postulates that efficient anchoring of subsequent to the tea1p–mod5p interaction at nongrowing tea1p at the cortex occurs in several stages, and ‘multimode’ cell tips, which accords with the demonstrated role for tea3p because we propose that the mechanisms of tea1p anchoring in the retention of tea1p at these sites. Furthermore, since are different in the presence or absence of mod5p, and at deletion of tea3p does not phenocopy tea1D200, we propose growing versus nongrowing cell tips. According to the model, that there may also be a ‘parallel mode’ of tea3p-independent tea1p is initially in a transport complex on microtubule plus tea1p retention at the growing cell tip (see below). ends, in association with other plus-end binding proteins (Figure 7A; also see below; Busch et al, 2004; Feierbach Parallel tea1p-anchoring systems at growing and nongrowing et al, 2004). Upon reaching cell tips, tea1p interacts, via its tips. Our results suggest that there are both common and N-terminus, with a central region of mod5p (Figure 7B). This distinct elements in the mechanisms regulating tea1p locali- interaction is important for the anchoring of tea1p at the zation at growing versus nongrowing cell tips. Whereas cortex (Snaith and Sawin, 2003) but is by itself insufficient mod5p and the tea1p C-terminus are each required for for the correct steady-state localization of tea1p and mod5p. tea1p anchoring at both cell tips, tea3p is required primarily Tea1p then further interacts, via its C-terminus, with corti- for tea1p anchoring at nongrowing tips, consistent with its cally localized tea3p. This may occur at both cell tips but preferential localization to nongrowing tips (Arellano et al, is functionally important primarily at nongrowing tips 2002). Thus, at growing cell tips, other tea1p-binding proteins (Figure 7C). Our data further indicate that tea3p is itself may perform functions analogous to those performed by anchored at cell tips by a strong interaction with a central tea3p at nongrowing tips. Further support for the suggestion region of mod5p (Figure 7D). We also suggest that at growing that a parallel pathway at growing cell tips involves the cell tips, there may be a parallel, tea3p-independent pathway tea1p–mod5p interaction comes from the observation that of tea1p retention, involving both a tea1p–mod5p interaction mod5D tea3D double mutants form ectopic branches at a and an interaction of tea1p with other cortical proteins much higher frequency than tea3D single mutants (Figure 7E). Finally, we propose that in addition to the (Figure 3A). This indicates that mod5p makes additional mod5p-dependent pathway of cortical tea1p anchoring, a contributions to polarity fidelity outside a strictly ‘tea3p- complementary mechanism exists whereby tea1p and tea3p specific’ pathway. Recently, a novel SH3-domain protein, functionally interact both in the cytoplasm and at the cortex tea4p/wsh3p, has been identified that may perform the to ‘bypass’ mod5p (Figure 7F). Tea1p anchored at cell tips ‘tea3p-analogous’ function at growing cell tips; interestingly, (Figure 7G) would then be capable of making further inter- tea4p, like tea3p, binds to the C-terminus of tea1p (Martin actions with its ‘downstream’ effectors, perhaps also further et al, 2005; Tatebe et al, 2005). strengthening its retention as a result. The tea1p–tea3p interaction can ‘bypass’ mod5p. This sug- Structure of the model gestion is based not only on the strong tea1p–tea3p inter- action seen in mod5D cells but also on mutant cell-polarity Because of the diverse nature of the results contributing to phenotypes. Both tea1D and tea3D mutants have more severe our model, we discuss more specific aspects of the model monopolar growth defects than mod5D (Figure 5B), indicat- below. ing that tea1p and tea3p still retain some function in organiz- A cortical interaction between tea3p and mod5p. Several ing growth polarity in the absence of mod5p, even if they are pieces of evidence support the idea that tea3p and mod5p not stably anchored at cell tips during steady-state growth. form a stable cortical complex. First, tea3p and mod5p have a Further support comes from the observation that mod5D tea3D double mutants form ectopic branches at a much robust biochemical interaction both in the presence and higher frequency than mod5D single mutants (Figure 5A). absence of tea1p; indeed, tea3p was the only strong interactor Moreover, antipodal growth persists in mod5D mutants as with mod5p identified in our two-hybrid screen. Second, we long as there is constant delivery of tea1p to cell tips by note that the delocalized tea3p seen in tea1D cells is never- theless still associated with the cell cortex, in a pattern microtubules (Snaith and Sawin, 2003). Most importantly, closely resembling that of mod5p (Figure 3; Arellano et al, our proposal for a ‘mod5p bypass’ is directly supported by the observation that tea3p is cotransported with tea1p to cell 2002). As this cortical localization is lost when both tea1p tips when mod5þ is deleted (Figure 4). and mod5p are deleted, it seems likely that the aberrant cortical localization of tea3p in tea1D mutants is due to its Limits of the model. An important issue not explicitly stable interaction with mod5p. addressed in the model is how the hypothesized positive The C-terminus of tea1p is important subsequent to the tea1p– feedback between tea1p and mod5p localization may operate (Snaith and Sawin, 2003). We have shown that two mod5 mod5p interaction. Tea1D200p is found at microtubule plus internal-deletion mutants fail to be restricted to cell tips even ends at both cell tips but is not properly anchored at the when tea1p is properly localized at the cortex, but we cannot cortex (Behrens and Nurse, 2002). We had initially expected yet distinguish whether this is directly due to their failure to this to be due to a failure of tea1D200p to bind to membrane- associated mod5p, but instead we found that tea1D200p can bind to tea1p, or to tea3p, or both. This will benefit from interact with mod5p in vivo. This suggests that the ‘failure further studies. point’ of tea1D200p in a tea1p-anchoring pathway occurs after the tea1p–mod5p interaction, and that interactions of Concluding remarks the tea1p C-terminus with other proteins are additionally In summary, a complex set of pairwise protein–protein inter- required for tea1p anchoring. Because tea1D200p fails to actions among tea1p, tea3p and mod5p is essential for the &2005 European Molecular Biology Organization The EMBO Journal VOL 24 NO 21 2005 3697 | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al Immunoprecipitation studies correct cortical anchoring of tea1p after its delivery to cell tips For immunoprecipitation experiments from yeast cell extracts, GST- by microtubules. In many eukaryotic systems, proteins asso- mod5p was expressed from the nmt41 promoter to steady-state ciated with microtubules are subsequently targeted to the levels. Tea1p, tea3p-GFP and tea3p-HA were all expressed from their cell cortex, and it is likely that this is achieved by multiple endogenous promoters. A 400–500 mg portion of frozen cell pellets was ground with a mortar and pestle under liquid nitrogen mechanisms (Schuyler and Pellman, 2001; Gundersen, and extracts were prepared in 20 mM NaHEPES pH 7.5, 50 mM 2002; Carvalho et al, 2003; Akhmanova and Hoogenraad, K-acetate, 200 mM NaCl, 1 mM EDTA, 2 mM MgCl and 0.2% Triton 2005; Watanabe et al, 2005). In relation to this, we note that a X-100. A 5–8 mg portion of total protein was used in each key point of our ‘mod5p bypass’ model is that the two immunoprecipitation in a final volume of 500ml. Samples were incubated with 5mg of anti-HA 12CA5, anti-GST or anti-tea1p proposed mechanisms of achieving a tea1p–tea3p interaction antibodies, and 25ml of protein A or protein G sepharose beads as at cell tips are not only complementary but also mutually appropriate, for 2–3 h at 51C. All samples were separated on 8% exclusive. Moreover, the exclusive aspect of their relationship SDS–PAGE and subjected to Western blotting. emerges from the properties of the system as a whole. In wild- type cells, tea1p interacts with tea3p primarily at the cell Physiology and immunofluorescence experiments cortex, because tea3p is sequestered there by mod5p. Immunostaining with anti-tea1p serum and anti-tubulin antibodies was exactly as described (Snaith and Sawin, 2003). Images of entire However, in, and only in, mod5D cells, tea3p is no longer cell volume were captured using a Leica TCS-SP confocal micro- similarly sequestered, and tea1p can therefore interact with scope (Leica Microsystems, Milton Keynes, UK). Fluorescence tea3p in the cytoplasm to form stable complexes that then quantitation of anti-tea1p immunofluorescence (n¼ 130 cells) was travel to the cortex. Importantly, even if the interaction of performed as described (Snaith and Sawin, 2003). For costaining tea3D cells for actin and tea1p, cells were grown in YES and fixed tea1p and tea3p in mod5D cells is less productive than in exactly as described (Sawin and Nurse, 1998), and then digested for wild-type cells (mod5D cells still exhibit some defects in 15 min at 371Cin 25mg/ml zymolyase 20-T. Images were captured polarized growth), both scenarios result in an interaction of using wide-field microscopy on a Nikon TE300 system (Snaith and tea1p and tea3p at the cell cortex, which would then allow Sawin, 2003). Polarity re-establishment experiments were carried tea1p to interact with effector binding partners. In the context out as described (Snaith and Sawin, 2003). In all experiments, 4200 cells were counted for each sample. Treatment with 50mg/ml of ‘systems cell biology’, this may reflect a kind of homeo- MBC was used to disrupt microtubules and 200mM latrunculin B static design principle by which desired outputs can be was used to disrupt actin filaments, during log-phase growth. achieved, even with some degree of fine-tuning, in the face of changing amounts of a critical system component, in this Live cell imaging case mod5p. All images were collected on the Nikon TE300 microscope system. Bright-field imaging of cell growth patterns was performed as described (Glynn et al, 2001). At least 130 cell divisions (i.e., growth Materials and methods patterns in at least 260 daughter cells) were analyzed for each strain. For fluorescence imaging, cells were grown at 251C in EMM General methods and imaged at room temperature. Expression of mutant GFP-mod5p S. pombe methods were as described (Moreno et al, 1991). was induced for 2 days at 321C in EMM plus 150 nM thiamine Molecular biology methods were essentially as described (Sam- before shifting to 251C for 4 h prior to imaging single planes. brook and Russell, 2001). Epitope-tagged mod5þ and tea3þ Z-series of tea3p-GFP still images were deconvolved using Deltavision strains were constructed by PCR-based gene targeting as described Softworx (Applied Precision) and projected into a single plane. (Ba¨hler et al, 1998). Tea3p-GFP, tea3D and tea1D200 strains were Z-series from doubly labeled tea3p-mCh tea1p-GFP or tea3p-mCh gifts from P Nurse, The Rockefeller University, NY. The two-hybrid GFP-atb2p strains were projected into a single plane without further library was a kind gift from T Nakamura, Osaka City University, manipulation. Further details of imaging conditions are given in Japan, and pRSETB-mCherry and pRSETB-tdTomato were gifts of Supplementary data. R Tsien, UCSD, CA. Complete lists of plasmids and strains used in this work appear as Supplementary data. Supplementary data Supplementary data are available at The EMBO Journal Online. Antibodies Anti-GSTantibodies were obtained from Amersham (Little Chalfont, UK); anti-HA 12CA5 antibodies were a gift from I Stancheva, Acknowledgements University of Edinburgh, UK, and anti-TAT1 hybridoma supernatant was a gift from K Gull, University of Oxford, UK (Woods et al, We thank M Masterton for invaluable help in constructing yeast 1989). Affinity-purified tea1p antibodies were used for Western strains expressing mod5p deletion proteins; K Gull, T Nakamura, blotting, and serum was used for immunofluorescence (Snaith and P Nurse, H Ohkura I Stancheva and R Tsien for sharing strains and Sawin, 2003). Anti-tea1p antibodies were purified against a 6-His reagents; A Merdes and I Davis for critical reading of the manu- fusion of tea1p amino acids 554–1147 (tea1p-C) expressed in script; S Martin for helpful discussion of our data; and F Philp bacteria (Mata and Nurse, 1997). Tea1p-C was isolated as inclusion for continuing encouragement. HAS is a Caledonian Research bodies, solubilized in SDS and coupled to Affigel 15 as described Foundation Post-doctoral Research Fellow and KES is a Wellcome (Sawin et al, 1992). Alexafluor-conjugated phalloidin and second- Trust Senior Research Fellow. This work was supported by the ary antibodies were obtained from Molecular Probes (Eugene, OR). Caledonian Research Foundation and the Wellcome Trust. References Adachi Y, Toda T, Niwa O, Yanagida M (1986) Differential expres- Arellano M, Niccoli T, Nurse P (2002) Tea3p is a cell end marker sions of essential and nonessential alpha-tubulin genes in activating polarized growth in Schizosaccharomyces pombe. Curr Schizosaccharomyces pombe. Mol Cell Biol 6: 2168–2178 Biol 12: 751–756 Adams J, Kelso R, Cooley L (2000) The kelch repeat superfamily of Ba¨hler J, Wu JQ, Longtine MS, Shah NG, McKenzie III A, Steever proteins: propellers of cell function. Trends Cell Biol 10: 17–24 AB, Wach A, Philippsen P, Pringle JR (1998) Heterologous Akhmanova A, Hoogenraad CC (2005) Microtubule plus-end-tracking modules for efficient and versatile PCR-based gene targeting in proteins: mechanisms and functions. Curr Opin Cell Biol 17: 47–54 Schizosaccharomyces pombe. Yeast 14: 943–951 3698 The EMBO Journal VOL 24 NO 21 2005 &2005 European Molecular Biology Organization | | Tea1p, mod5p and tea3p in fission yeast HA Snaith et al Behrens R, Nurse P (2002) Roles of fission yeast tea1p in the Mitchison JM, Nurse P (1985) Growth in cell length in the fission localization of polarity factors and in organizing the microtubular yeast Schizosaccharomyces pombe. J Cell Sci 75: 357–376 cytoskeleton. J Cell Biol 157: 783–793 Moreno S, Klar A, Nurse P (1991) Molecular analysis of the fission Bretscher A, Edwards K, Fehon RG (2002) ERM proteins and merlin: yeast Schizosaccharomyces pombe. Methods Enzymol 194: integrators at the cell cortex. Nat Rev Mol Cell Biol 3: 586–599 795–823 Browning H, Hackney DD, Nurse P (2003) Targeted movement of Niccoli T, Arellano M, Nurse P (2003) Role of Tea1p, Tea3p and cell end factors in fission yeast. Nat Cell Biol 5: 812–818 Pom1p in the determination of cell ends in Schizosaccharomyces Browning H, Hayles J, Mata J, Aveline L, Nurse P, McIntosh JR pombe. Yeast 20: 1349–1358 (2000) Tea2p is a kinesin-like protein required to generate polar- Prag S, Adams JC (2003) Molecular phylogeny of the kelch-repeat ized growth in fission yeast. J Cell Biol 151: 15–28 superfamily reveals an expansion of BTB/kelch proteins in Brunner D, Nurse P (2000) CLIP170-like tip1p spatially organizes animals. BMC Bioinformatics 4: 42 microtubular dynamics in fission yeast. Cell 102: 695–704 Sambrook J, Russell DW (2001) Molecular Cloning: A Laboratory Busch KE, Hayles J, Nurse P, Brunner D (2004) Tea2p kinesin is Manual. New York: Cold Spring Harbor Laboratory Press involved in spatial microtubule organization by transporting Sawin KE, Mitchison TJ, Wordeman LG (1992) Evidence for tip1p on microtubules. Dev Cell 6: 831–843 kinesin-related proteins in the mitotic apparatus using peptide Carvalho P, Tirnauer JS, Pellman D (2003) Surfing on microtubule antibodies. J Cell Sci 101 (Part 2): 303–313 ends. Trends Cell Biol 13: 229–237 Sawin KE, Nurse P (1998) Regulation of cell polarity by microtu- Chang F, Peter M (2003) Yeasts make their mark. Nat Cell Biol 5: bules in fission yeast. J Cell Biol 142: 457–471 294–299 Sawin KE, Snaith HA (2004) Role of microtubules and tea1p Drubin DG, Nelson WJ (1996) Origins of cell polarity. Cell 84: in establishment and maintenance of fission yeast cell polarity. 335–344 J Cell Sci 117: 689–700 Dyson HJ, Wright PE (2005) Intrinsically unstructured proteins and Schuyler SC, Pellman D (2001) Microtubule ‘plus-end-tracking their functions. Nat Rev Mol Cell Biol 6: 197–208 proteins’: the end is just the beginning. Cell 105: 421–424 Feierbach B, Chang F (2001) Roles of the fission yeast formin for3p Shaner NC, Campbell RE, Steinbach PA, Giepmans BN, Palmer AE, in cell polarity, actin cable formation and symmetric cell division. Tsien RY (2004) Improved monomeric red, orange and yellow Curr Biol 11: 1656–1665 fluorescent proteins derived from Discosoma sp. red fluorescent Feierbach B, Verde F, Chang F (2004) Regulation of a formin protein. Nat Biotechnol 22: 1567–1572 complex by the microtubule plus end protein tea1p. J Cell Biol Small JV, Kaverina I (2003) Microtubules meet substrate adhesions 165: 697–707 to arrange cell polarity. Curr Opin Cell Biol 15: 40–47 Glynn JM, Lustig RJ, Berlin A, Chang F (2001) Role of bud6p and Snaith HA, Sawin KE (2003) Fission yeast mod5p regulates polar- tea1p in the interaction between actin and microtubules for the ized growth through anchoring of tea1p at cell tips. Nature 423: establishment of cell polarity in fission yeast. Curr Biol 11: 836–845 647–651 Gundersen GG (2002) Microtubule capture: IQGAP and CLIP-170 Snaith HA, Sawin KE (2005) Tea for three: control of fission yeast expand the repertoire. Curr Biol 12: R645–R647 polarity. Nat Cell Biol 7: 450–451 Hayles J, Nurse P (2001) A journey into space. Nat Rev Mol Cell Biol Tatebe H, Shimada K, Uzawa S, Morigasaki S, Shiozaki K (2005) 2: 647–656 Wsh3/Tea4 is a novel cell-end factor essential for bipolar dis- Jin H, Amberg DC (2001) Fission yeast Aip3p (spAip3p) is required tribution of Tea1 and protects cell polarity under environmental for an alternative actin-directed polarity program. Mol Biol Cell stress in S. pombe. Curr Biol 15: 1006–1015 12: 1275–1291 Vega LR, Solomon F (1997) Microtubule function in morphological Kim H, Yang P, Catanuto P, Verde F, Lai H, Du H, Chang F, Marcus S differentiation: growth zones and growth cones. Cell 89: 825–828 (2003) The kelch repeat protein, Tea1, is a potential substrate Verde F (2001) Cell polarity: a tale of two Ts. Curr Biol 11: target of the p21-activated kinase, Shk1, in the fission yeast, R600–R602 Schizosaccharomyces pombe. J Biol Chem 278: 30074–30082 Verde F, Mata J, Nurse P (1995) Fission yeast cell morphogenesis: Li X, Zhang D, Hannink M, Beamer LJ (2004) Crystal structure identification of new genes and analysis of their role during the of the Kelch domain of human Keap1. J Biol Chem 279: cell cycle. J Cell Biol 131: 1529–1538 54750–54758 Watanabe T, Noritake J, Kaibuchi K (2005) Regulation of micro- Martin SG, Chang F (2003) Cell polarity: a new mod(e) of anchor- tubules in cell migration. Trends Cell Biol 15: 76–83 ing. Curr Biol 13: R711–R713 Woods A, Sherwin T, Sasse R, MacRae TH, Baines AJ, Gull K (1989) Martin SG, McDonald WH, Yates III JR, Chang F (2005) Tea4p links Definition of individual components within the cytoskeleton microtubule plus ends with the formin for3p in the establishment of Trypanosoma brucei by a library of monoclonal antibodies. of cell polarity. Dev Cell 8: 479–491 J Cell Sci 93: 491–500 Mata J, Nurse P (1997) Tea1p and the microtubular cytoskeleton are Yarm F, Sagot I, Pellman D (2001) The social life of actin important for generating global spatial order within the fission and microtubules: interaction versus cooperation. Curr Opin yeast cell. Cell 89: 939–949 Microbiol 4: 696–702 &2005 European Molecular Biology Organization The EMBO Journal VOL 24 NO 21 2005 3699 | |

Journal

The EMBO Journal – Springer Journals

Published: Nov 2, 2005

Keywords: cell polarity; kelch repeat; mCherry; S. pombe; tdTomato

There are no references for this article.