Rgf1 GEF activity toward Rho1 defines a new actin-dependent signal to determine growth sites independently of microtubules and Tea1

1 1 2 3 Rgf1 GEF activity toward Rho1 defines a new actin-dependent signal to determine growth sites 4 independently of microtubules and Tea1 5 6 7 8 Patricia Garcia*, Ruben Celador, Tomas Edreira and Yolanda Sanchez* 9 10 Instituto de Biología Funcional y Genómica (IBFG), CSIC/Universidad de Salamanca and 11 Departamento de Microbiología y Genética, Universidad de Salamanca. C/ Zacarías González, 12 s/n. 37007 Salamanca, Spain. 13 14 Short title: Rgf1 and actin define growth sites 15 * Correspondence should be addressed to: [email protected] ; [email protected] 16 Key words: polarity, actin, Rho-GTPases, yeast 17 18 19 20 21 22 23 24 25 26 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 2 27 ABSTRACT 28 Cellular asymmetry begins with the selection of a discrete point on the cell surface that triggers 29 Rho-GTPases activation and localized assembly of the cytoskeleton to establish new growth 30 zones. The cylindrical shape of fission yeast is organized by microtubules that deliver the 31 landmark Tea1–Tea4 complex at the cell tips to define the growth poles. However, only a few 32 tea1Δ cells mistaken the direction of growth, indicating that they manage to detect their growth 33 sites. Here we show that Rgf1 (Rho1-GEF) and Tea4 are components of the same complex and 34 that Rgf1 activity toward Rho1 is required for strengthen Tea4 at the cell tips. Moreover, in cells 35 lacking Tea1, selection of the correct growth site depends on Rgf1 and on a correctly polarized 36 actin cytoskeleton, both necessary for Rho1 activation at the pole. We propose an actin- 37 dependent mechanism driven by Rgf1–Rho1 that marks the poles independently of 38 microtubules and the Tea1–Tea4 complex. 39 40 41 42 43 44 45 46 47 48 49 50 51 52 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 3 53 INTRODUCTION 54 Cell polarity is the primary mechanism for generating cellular asymmetry, which is critical for 55 most cell and tissue functions such as development, cell migration and differentiation in a wide 56 variety of organisms including humans. It typically begins with a signal on the cell surface that 57 triggers a cascade of molecular events that induce the localized assembly of cytoskeletal and 58 signaling networks, which subsequently direct the formation of a new growth area (1,2). Fission 59 yeast has been used over the last decades as a simpler and more accessible model for studying 60 this complex process (3). Its cells have a cylindrical shape that is maintained throughout the 61 entire cell cycle, changing in length but not in diameter. This phenomenon is achieved by 62 restricting growth to the cell poles, a process that is still not well understood. Growth occurring 63 at the cell ends has been mainly studied in the transition from monopolar to bipolar growth, 64 termed New End Take-Off (NETO) (4–7). NETO depends on specific polarity determinants, the 65 kelch-repeat protein Tea1, the SH3 domain-containing protein Tea4 and the DYRK kinase, Pom1 66 among others (8–11). In the absence of Tea1–Tea4 complex, cells grow monopolarly but 67 maintain their cylindrical shape. However, under certain stresses, the cells often choose the 68 wrong growth site, forming bulged and T-shaped cells (10,12). 69 Tea1 and Tea4 ride on growing microtubule (MT) plus ends to the cell tips, where they 70 are released as discrete “dots” at the cortex, being Tea4 totally dependent on Tea1 for its 71 location (12–15). At poles, the prenylated protein Mod5 and the ERM (Ezrin-Radixin-Moesin) 72 family protein Tea3 anchor Tea1 to the cell cortex (16,17). Tea1 and Tea4 colocalize at the cell 73 tips to form clusters or nodes, (18) and it is assumed that this association promotes the binding 74 of other polarity factors in large protein complexes that organize polarized growth (4,10,19). 75 One of these proteins is the formin For3, whose association with polarity markers likely brings 76 it into the proximity of activators, stimulating the formation of F-actin cables that will deliver 77 growth cargo to the tip (9). .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 4 78 Establishing polarized growth involves hundreds of proteins; however, a constant from 79 yeast to humans involves local accumulation of active GTP-bound forms of Rho-family GTPases 80 at the cell cortex (20–22). While polarity is mostly associated with the functions of Rac and Cdc42 81 in mammals or Cdc42 in yeast, other GTPases such as RhoA or Rho1 (yeast) can also play a role 82 in the development of polarity. Active RhoA is found at the leading edge as the edge advances 83 in migrating cells, whereas Cdc42 and Rac1 are activated later (23). In budding and fission yeast, 84 Cdc42 displays the ability to polarize spontaneously (22,24–26). However, the role of Rho1, the 85 other essential GTPase, in polarized growth remains undefined. Depletion of Rho1 in growing 86 cells induces shrinking and death via a kind of “apoptosis” that is accompanied by the 87 disappearance of polymerized actin (27,28). Rho1 activity is regulated by three GEFs, Rgf1, Rgf2, 88 and Rgf3 that catalyze the exchange of GDP for GTP, rendering the GTPase in an active state 89 (28–34). The main activator of Rho1, Rgf1, regulates cell integrity through Rho1 by activating the 90 β-glucan-synthase complex (28) and gene expression via the Pmk1 MAPK cell integrity-signaling 91 pathway (28,35,36). Moreover, Rgf1, like Tea1, Tea4 and other polarity factors, is required for 92 the actin reorganization necessary to switch from monopolar to bipolar growth during NETO 93 (28). 94 Here we have studied this phenomenon to show that Rgf1 interacts with the cell end 95 marker Tea4 and binds to the plasma membrane (PM) through its PH domain. Both, PM binding 96 and Rho-GEF activity are required for stable accumulation of polarity markers at the cell poles. 97 In addition, we have uncovered a new role for Rgf1 in restricting growth to the poles in the 98 absence of polarity markers. Most tea1Δ cells maintain their cylindrical morphology unless 99 subjected to stresses, suggesting that these cells detect the location of its poles by an unknown 100 mechanism. Here we show that this mechanism depends on the actin cytoskeleton and Rho1 101 activation by Rgf1. Therefore, we propose two parallel pathways to define the growth poles in 102 fission yeast: the canonical one dependent on MTs and Tea1–Tea4 and another one dependent 103 on actin and Rgf1–Rho1, both necessary to maintain a straight shape when the other is impaired. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 5 104 105 106 RESULTS 107 Rgf1 is required for proper localization of the Tea1–Tea4 complex 108 We have previously shown that rgf1Δ strain displays defects in bipolar growth, with ~80% of the 109 cells showing monopolar pattern of growth compared to ~25% of the wild-type cells (28). This 110 growth defect has been described in mutants affected in polarity factors such as Bud6, Tea1 or 111 For3 (37–39), suggesting that similarly to these proteins, Rgf1 triggers NETO, and thus rgf1Δ cells 112 may have problems when choosing the right end of growth. This prompted us to examine the 113 localization of the landmark proteins Tea4 and Tea1 in cells lacking Rgf1. While Tea4-GFP was 114 concentrated at both cell tips in wild-type cells (9,11), in rgf1Δ cells the signal detected was 115 visibly diminished. We compared the fluorescence intensity of Tea4-GFP in wild-type and rgf1Δ 116 cells in the same preparation, in which we grew tea4-GFP rgf1+ sad1-dsred (spindle pole body 117 marker) cells and tea4-GFP rgf1Δ cells separately, and then mixed and imaged them at the same 118 time (Fig 1A). In rgf1Δ cells, the Tea4-GFP signal was dispersed in small dots that spread out at 119 the ends. The average fluorescence of Tea4-GFP dots at the tips of rgf1Δ cells was approximately 120 half of that seen at the tips of the wild-type cells. Because the rgf1Δ cells grew in a monopolar 121 fashion, we examined whether the Tea4 dots were more prominent at one end or whether they 122 were scattered similarly at both ends. We observed that 61% of the rgf1Δ cells accumulated 123 Tea4 at the non-growing end (revealed by calcofluor staining) (Fig 1B). Thus, Rgf1 is more 124 important to localize Tea4 to the growing tip, the one where Rgf1 concentrates in wild-type 125 monopolar cells (Fig S1A). 126 127 Fig 1: Rgf1 is required for proper localization of the Tea1-Tea4 complex at the cell tip 128 (A) Cells expressing tea4-GFP rgf1 + sad1-dsred and tea4-GFP rgf1Δ were grown in YES liquid 129 medium separately, and then mixed and imaged in the same preparation. The maximum- .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 6 130 intensity projection of six deconvolved Z-slides (0.5 µm step-size) is shown. The graphic 131 represents the mean ± standard deviation (SD) of the relative fluorescence intensity of Tea4- 132 GFP measured at the cell tips in the wild-type (WT, n = 47) and rgf1Δ (n = 41) cells. WT levels 133 were used for normalization. (B) Calcofluor white (CW, 20 µg/mL) staining and GFP fluorescence 134 in cells expressing tea4-GFP rgf1+ and tea4-GFP rgf1Δ. The maximum-intensity projection of six 135 Z-slides (0.5 µm step-size) of Tea4-GFP fluorescence is shown. The red arrowheads indicate non- 136 growing poles. The graphic represents the mean ± SD of the relative fluorescence intensity of 137 Tea4-GFP measured at the tips of the rgf1Δ (n = 50) cells. Calcofluor staining, which marks sites 138 of cell growth, was used to differentiate growth from non-growth poles (right). (C) 139 Representative images of Tea4-GFP (green) and Tea1-tdTomato (red) localization in the rgf1Δ 140 cells. The maximum-intensity projection of six Z-slides (0.5 µm step-size) is shown. (D) 141 Representative images of the indicated cells expressing tea4-GFP (green) and mCherry-atb2 142 (red). The maximum-intensity projection of six Z-slides (0.5 µm step-size) is shown. (E) 143 Kymographs of time-lapse fluorescence movies of tea4-GFP and mCherry-atb2 expressed in 144 wild-type or rgf1Δ strains. The maximum-intensity projection of seven Z-slides (0.6 µm step-size) 145 of images taken every 15 s was used to draw a line along an MT from the middle of the cell to 146 the tip. The orange arrowheads indicate the moment of MT retraction. (F) Super-Resolution 147 Radial Fluctuations (SRRF) images of Tea4-GFP (green) and mCherry-Atb2 (red) in “head-on” cell 148 tips of the indicated strains. One focal plane image was taken every minute. The time projection 149 of the three images at different time points is shown to follow Tea4 cluster movement (right). 150 Note that in the wild-type strain Tea4 nodes remain stable (white arrowheads) for longer than 151 in the rgf1Δ mutant (blue arrowheads). (G) Kymographs of time-lapse fluorescence movies of 152 GFP-Atb2 producing in the WT and rgf1Δ cells. The graph shows the mean ± SD of the time during 153 which the MT is touching the pole in both strains (n = 75). (H) Representative images of the 154 indicated cells producing mCherry-Atb2. The maximum-intensity projection of six Z-slides (0.5 155 µm step-size) is shown. The graph shows the mean ± SD of the percentage of curved MTs found .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 7 156 in the indicated strains (n > 500). Statistical significance was calculated using two-tailed unpaired 157 Student’s t test. ****P < 0.0001; ns = non-significant. Scale bar 2 µm. 158 159 We also evaluated whether the localization of Tea1, which functions in a complex with 160 Tea4 (9), is affected in the Rgf1 mutant. The wide co-localization between Tea1 and Tea4 in the 161 absence of Rgf1 indicates that both proteins displayed similar localization defects (Fig 1C), and 162 that similarly to wild-type cells, in rgf1Δ cells Tea1 and Tea4 were still bound and formed a 163 complex. In tea4Δ cells, Tea1 is concentrated at the non-growing cell tip (9). Given that in rgf1Δ 164 cells, Tea1 also localized mainly to the non-growing end (Fig 1B and C), our results suggest that 165 the Tea1 mislocalization could be a consequence of Tea4 mislocalization. We noticed that rgf1Δ 166 cells showed a greater number of Tea4 discrete dots in the middle zone of the cell (Figs 1A, 1B, 167 and S1B). Examination of Tea4-GFP in wild-type and rgf1Δ cells with α–tubulin labeled in red 168 (mCherry-Atb2) showed co-localization of Tea4 cytoplasmic dots with MTs (Fig 1D). Thus, in the 169 rgf1Δ cells, a larger free cytoplasmic pool of Tea4, which is not properly sequestered at the poles, 170 could now be available to be redirected to the cell cortex by MTs. 171 172 Rgf1 functions to integrate Tea4 in big clusters at the cell tip 173 Next, we determined whether Tea4 was accurately delivered to the cell cortex in rgf1Δ cells by 174 taking time-lapse images every 15 seconds. We did not observe appreciable differences in the 175 delivery of Tea4 to the cell cortex between rgf1+ and rgf1Δ cells. However, the Tea4-GFP signal 176 failed to remain in the pole in rgf1Δ cells (Movies S1 and S2). This can be better observed in the 177 kymographs shown in Fig 1E, where the fluorescence of Tea4 vanished from the cell cortex of 178 the rgf1Δ cells in a few seconds, whereas it remained stable in control cells. 179 Polarity factors such as Tea1 and Tea4 localize to the cellular cortex in discrete clusters 180 (18), which are not easily observable when taking conventional lateral cell images. To better 181 perceive the formation of Tea4 nodes at the cell poles we used Super-Resolution Radial .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 8 182 Fluctuations (SRRF) microscopy for “head-on” imaging of cell tips. We performed short time- 183 lapse experiments (three minutes) at the cellular tip cortex on “head-on” wild-type and rgf1Δ 184 cells (tagged with Tea4-GFP and mCherry-Atb2). In wild-type cells, Tea4 nodes deposited by MTs 185 remained stable for an interval of time even after MT catastrophe (Fig 1F, white arrows). 186 Moreover, clusters of Tea4 not associated with microtubules could be observed stable 187 throughout the time-lapse, and were of the same size as those associated with MTs. However, 188 in rgf1Δ cells, Tea4 dots of a similar size to those observed in the wild-type cells appeared 189 exclusively while they were associated with MTs (Fig 1F, blue arrows). Once the MT retracted, 190 the Tea4 cluster became gradually smaller until it eventually disappeared. These observations 191 confirmed the results obtained with conventional microscopy methods, suggesting that Rgf1 192 was not necessary for the delivery of Tea4 to the cell poles, but it is required for its stable 193 maintenance once it was released there. In addition, we ruled out a defect in the stability of the 194 Tea4 protein in the rgf1Δ mutant because the half-life of the protein in the wild-type and rgf1Δ 195 cultures treated with cycloheximide was comparable (Fig S1C). 196 During the course of these experiments, we noticed that some MTs curled around the 197 tips in the rgf1Δ cells. We confirmed that the MT dynamics was not affected because the 198 polymerization and depolymerization rates were similar in the wild-type and rgf1Δ cells, with a 199 slight increase in the polymerization rate in the mutant (Fig S1D). However, the mean time that 200 the MT stayed at the tip was ~43 seconds in the rgf1+ cells but ~70 seconds in the rgf1Δ cells (Fig 201 1G). Therefore, once an MT reached the cortex in the rgf1Δ cells, it remained there longer than 202 in the wild-type cells. Curved MTs have already been described in tea1Δ cells growing at a high 203 temperature (10). We incubated the rgf1Δ and tea1Δ mutants at 37°C for 4 hours and observed 204 MT organization under these conditions (Fig 1H). In the rgf1Δ mutant, ~10% of the cells 205 possessed at least one MT curled around the end, similarly to the ~12% found in the tea1Δ 206 mutant, while this type of curly MT was rarely observed in the wild-type (1.7%). It is possible .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 9 207 that a limited amount of Tea4–Tea1 at the growing pole underlies the curly phenotype seen in 208 the absence of Rgf1. 209 210 Rgf1 cooperates with Mod5 in Tea4 anchoring to the cellular poles. 211 It has been shown that in cells lacking Mod5, Tea1 and Tea4 fail to accumulate to wild-type levels 212 at the cell tips (9,16,40). Given that the rgf1Δ cells showed a similar defect, we analyzed the 213 localization of Tea4 in the double mutant rgf1Δmod5Δ compared with the single mutants. In the 214 mod5Δ and rgf1Δ cells, most of the Tea4-GFP dots were associated with MTs, although was still 215 some signal mainly at one of the poles (Fig 2A, yellow arrows ). However, in the rgf1Δmod5Δ 216 double mutant, the Tea4 signal was depleted at both poles (Fig 2A, orange arrows), suggesting 217 that Mod5 and Rgf1 share a function in Tea4 anchoring. 218 219 Fig 2: Rgf1 cooperates with Mod5 in Tea4 anchoring to the cellular poles 220 (A) Representative images of the indicated strains producing Tea4-GFP (green) and mCherry- 221 Atb2 (red). The maximum-intensity projection of six Z-slides (0.5 µm step-size) is shown. Scale 222 bar 2 µm. (B) The percentage of cells forming branches in the indicated strains. The cells were 223 grown to the stationary phase for 3 days, and then were treated with DMSO (-MBC) or MBC 224 (+MBC; 50 µg/ml). The mean ± SD of > 200 cells from three independent experiments is shown. 225 (C) The percentage of T-shaped cells in the indicated strains after incubation for 4 hours at 36°C 226 in YES liquid medium. Statistical significance was calculated using a two-tailed unpaired 227 Student’s t test. ns = non-significant. 228 229 Because the localization of Tea4 is entirely dependent on Tea1 and the localization of 230 Tea1 is partially dependent on Tea4 (9), we wondered whether the phenotype of the 231 rgf1Δmod5Δ cells (which mislocalize Tea4) resemble the tea1Δ phenotype. To this end, we 232 scored the percentage of T-shaped cells in polarity re-establishment assays. In these assays, we .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 10 233 first grew cells to the stationary phase for 3 days, and then diluted them in fresh medium for 3 234 hours. This treatment increases the penetrance of polarity mutant phenotypes, inducing T- 235 shapes and bulges and was performed either in the presence or in the absence of the MT 236 inhibitor methyl-2-benzimidazole carbamate (MBC), to prevent or allow the continuous delivery 237 of Tea1–Tea4 to the poles by microtubules(12,16). Consistent with the “poor” localization of 238 Tea4 in the cell cortex, the rgf1Δ and mod5Δ cells showed polarity defects when treated with 239 MBC and marginal defects in the absence of MBC, whereas ~80% of the tea1Δ cells displayed 240 the characteristic T-shaped pattern, with or without MBC treatment (Fig 2B and S2A )(16). 241 Interestingly, the rgf1Δmod5Δ cells behaved similar to the tea1Δ mutant: reaching ~80% of T- 242 shaped cells with MBC and showing ~45% even in the presence of MT (-MBC) (Fig 2B and S2A). 243 In addition, we combined the rgf1Δmod5Δ deletion with the temperature sensitive (ts) septation 244 mutant cdc11-119. At the restrictive temperature, cdc11-119 cells show a defect in cytokinesis, 245 but the nuclear and growth cycles continue and cells grow at both ends after each mitosis. 246 Presumably, after each mitosis, cdc11-119 mutant cells must decide where to reinitiate growth. 247 In the cdc11-119 tea1Δ double mutant, these events are often aberrant, leading to the 248 formation of highly branched or T-shaped multinuclei cells (10,41). Only ~5% of the cdc11rgf1Δ 249 and cdc11mod5Δ cells were T-shaped after incubation for 4 hours at 36°C ( Fig 2C and S2B). 250 However, the cdc11rgf1Δmod5Δ triple mutant showed a similar percentage of T-shaped cells 251 (~17%) as the cdc11tea1Δ double mutant (~21%). These results indicate that Rgf1 and Mod5 252 collaborate to position the polarity markers at the cell tips to prevent mislocalization of growth 253 machinery in successive cell cycles. 254 255 Rgf1 interacts with the cell-end marker Tea4 and binds to phosphatidylinositol-4-phosphate 256 through its PH domain. 257 Tea1 and Tea4 reside in large protein complexes (10,19). We used different approaches to 258 determine whether Rgf1 acts locally to retain Tea4 at the cell tips. First, we examined the in vivo .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 11 259 localization of endogenous Tea4-GFP together with Rgf1-tdTomato. In newly divided and 260 interphase cells, a subset of Rgf1 dots colocalized with Tea4 dots at cell tips, indicating a close 261 proximity with each other (Fig 3A). Subsequently, we tested for the coprecipitation of these two 262 proteins from yeast extracts by using epitope-tagged strains. Indeed, endogenously expressed 263 GFP-tagged Tea4 led to the co-purification of HA-tagged Rgf1(Fig 3B). Tea1 forms a stable 264 complex with Tea4 (9); however, we could not detect the interaction between Tea1 and Rgf1 265 (Fig S3). To validate these associations, we purified GST-Rgf1 from Escherichia coli and 266 conjugated it with Glutathione Sepharose (GS) beads; then, we used those beads in a pull-down 267 assay to trap Tea4-GFP or Tea1-GFP from Shizosaccharomyces pombe protein extracts. We 268 detected binding of Tea4 when using Rgf1-GS beads but not with GS beads alone (Fig 3C). In 269 addition, we could observe a slight precipitation of Tea1, which might be the result of Tea4 270 interaction with Tea1. We confirmed the biochemical interaction in a two-hybrid assay. 271 Consistently, Rgf1 could interact with Tea4 but not with Tea1 (Fig 3D). Taken together, these 272 results indicate that Rgf1 associates with the Tea1–Tea4 complex through its binding with Tea4. 273 274 Fig 3: Rgf1 interacts with the cell end marker Tea4 and binds to phosphatidylinositol-4- 275 phosphate through its PH domain. 276 (A) Colocalization of Rgf1 and Tea4. Representative images of wild-type cells producing Tea4- 277 GFP endogenously (green) and Rgf1-tdTomato from a plasmid under the control of its own 278 promoter (red). The maximum-intensity projection of six Z-slides (0.5 µm step-size) is shown. (B) 279 Coprecipitation of Rgf1 and Tea4. Cell extracts from cells producing Tea4-GFP, Rgf1-HA, or Tea4- 280 GFP and Rgf1-HA were precipitated with GFP-trap beads and blotted with anti-HA or anti-GFP 281 antibodies (co-immunoprecipitation and immunoprecipitation). Western blot was performed on 282 total extracts to visualize total Tea4-GFP and Rgf1-HA levels (whole cell extracts). (C) Cells 283 expressing Tea1-GFP or Tea4-GFP were pulled down from cell extracts with GST-Rgf1 purified 284 from E. coli bound to GS-beads or with GS-beads alone, and blotted with anti-GST or anti-GFP .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 12 285 antibodies (Pull-down). Total Tea1-GFP or Tea4-GFP levels (WCE) were visualized by western 286 blot; tubulin was used as a loading control. (D) Two-hybrid analysis of the interaction between 287 Tea1 (pGR135) and Tea4 (pGR106) with Rgf1 (pRZ97). The interaction was assessed on YNB 288 plates without histidine (YNB -H). (E) Protein-lipid overlay assay. Membrane lipid strips were 289 overlaid with 1 ug/ml of the purified GST, GST-Rgf1, and GST-Rgf1ΔPH respectively, and the 290 interaction was detected with an anti-GST antibody. Lipids to which GST-Rgf1 showed a 291 significant association are shown in red. Note that the interaction with PI4P disappears with 292 GST-Rgf1ΔPH. (F) Representative images of cells producing Rgf1-GFP or Rgf1ΔPH-GFP. The 293 maximum-intensity projection of six Z-slides (0.5 µm step-size) is shown. The graphic represents 294 the mean ± SD of the relative fluorescence intensity measured at the cellular tips of Rgf1-GFP 295 and Rgf1ΔPH-GFP (n>120). (G) Representative images of cells producing Rgf1-GFP in WT or efr3Δ 296 mutant. The maximum-intensity projection of six Z-slides (0.5 µm step-size) is shown. The 297 graphic represents the mean ± SD of the relative fluorescence intensity measured at the cellular 298 tips of Rgf1-GFP (n > 120). (H) Protein extracts from cell producing Rgf1-GFP in the WT or efr3Δ 299 mutant and Rgf1ΔPH-GFP were analyzed by western blot with an anti-GFP antibody to visualize 300 Rgf1 levels. An anti-tubulin antibody was used as a loading control. The graphic represents the 301 mean ± SD of the relative Rgf1 proteins levels from two independent experiments. Statistical 302 significance was calculated using a two-tailed unpaired Student’s t test. ****P< 0.0001; ***P< 303 0.001; **P < 0.01. Scale bar 2 µm. 304 305 Next, we evaluated whether Rgf1 acts as a linker between Tea4 and the PM in the 306 anchoring process. Rgf1 is a large (~150 KDa) multi-domain protein, including a pleckstrin 307 homology (PH) domain (34). PH domains could act as a “membrane-targeting device” by 308 anchoring GEFs to phosphoinositides and directing them towards their partner GTPases on the 309 cellular cortex (42,43). To test the ability of Rgf1 to bind different membrane lipids, we fused 310 the Rgf1 protein to GST (without its carboxi-terminal CNH domain to purify it more easily) and .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 13 311 purified it from bacteria. We used membrane arrays spotted with different kind of lipids to 312 detect binding of GST-Rgf1 to lipids. As shown in Fig 3E, recombinant-purified GST-Rgf1 313 preferentially bound to phosphatidic acid (PA) and cardiolipin and also bound to 314 phosphatidylinositol 4-phosphate [PI(4)P], and 3-sulfogalactosylceramide more subtly. GST- 315 Rgf1∆PH (additionally lacking the PH domain) could not interact with PI(4)P, but it behaved like 316 the wild-type protein in terms of its binding to the other lipids. This result contrasts with that 317 described for the budding yeast Rgf1 homolog, Rom2, whose PH domain specifically binds to 318 phosphatidylinositol 4,5-bisphosphate [PI(4,5)P2] (44). Thus, in S. pombe Rgf1 might bind to the 319 PM through the interaction between its PH domain with the phospholipid PI(4)P. Consistently, 320 the PH domain was required for proper anchoring of Rgf1 to the cellular cortex. A GFP-tagged 321 Rgf1∆PH mutant showed large defects in its localization at both the poles and the septum (Fig 322 3F). The fluorescence detected at the rgf1∆PH-GFP cell ends was ~25% of that exhibited by the 323 full-length protein Rgf1-GFP (Fig 3F, right). 324 Kathleen Gould’s group previously reported that Rgf1 displays septum localization 325 defects in cells lacking the gene efr3+, a PM scaffold for the PI(4)P kinase Stt4. Cells lacking efr3+ 326 do not properly position Stt4 and display altered levels of PM phosphoinositides (45). We found 327 that the localization of Rgf1 at the cell poles was compromised in the efr3∆ cells ( Fig 3G). 328 Moreover, the PM binding of Rgf1 might be crucial for protein stability. We detected a 3–4-fold 329 decrease in the protein level of Rgf1, both without the PH domain or in efr3∆ cells, when Rgf1 330 was not properly bound to the PM (Fig 3H). These results support a role for PM phospholipids 331 in the anchoring of Rgf1 to the cellular cortex and in the maintenance of Rgf1 protein level. 332 333 Tea4 accumulation at the cell ends depends on Rgf1 anchoring to the PM and Rho1 activation. 334 Next, we determined whether Tea4 is bound to the cell cortex in the rgf1∆PH-GFP mutant, which 335 shows compromised Rgf1 localization. Cortical localization of Tea4 was greatly reduced (~35%) 336 in the mutant lacking the PH domain compared with the wild-type (P < 0.0001) (Fig 4A). .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 14 337 However, in the rgf1∆PH-GFP mutant the ability to join and anchor Tea4 to the cortex could be 338 reduced due to a significant drop in the protein level compared with the Rgf1-GFP protein (Fig 339 3H). GTPase activation by its GEFs usually takes place when both the GTPase and the GEF are 340 close to the PM; thus, a low level of Rgf1∆PH-GFP at the PM could promote inefficient activation 341 of the Rho1 GTPase. To examine whether this is the case, we analyzed the in vivo amount of 342 GTP-Rho1 (active-Rho1) in the rgf1∆PH cells in a pull-down assay with GST-C21RBD, the 343 rhotekin-binding domain (previously purified from bacteria) (28). We found that the level of 344 active-Rho1 detected in the rgf1∆PH cells was similar to that seen in the rgf1∆ cells, and much 345 less than the amount detected in control cells (Fig 4B). Both, the localization of Rgf1 to the PM 346 and the activation of Rho1 were impaired in the rgf1∆PH mutant; hence, we could not determine 347 which one is behind Tea4 mislocalization. To distinguish between these two possibilities, we 348 utilized the rgf1-ΔPTTR mutant expressing a protein without four amino acids in the RhoGEF 349 domain. This mutant displays significantly reduced GEF activity toward Rho1,(36) while the 350 protein remained attached to the growing end (Fig S4). Tea4-GFP was also mislocalized in the 351 rgf1-ΔPTTR mutant (Fig 4A), indicating that the stable association of Tea4 to the membrane is 352 dependent on Rgf1 GEF activity. 353 354 Fig 4: Tea4 accumulation at the cell ends depends on Rgf1 anchoring to the PM and Rho1 355 activation. 356 (A) Representative images of Tea4-GFP localization in the WT, rgf1Δ, rgf1ΔPH, and rgf1ΔPTTR 357 cells. The maximum-intensity projection of six Z-slides (0.5 µm step-size) of Tea4-GFP 358 fluorescence is shown. Scale bar 2 µm. The graphic represents the mean ± SD of the relative 359 fluorescence intensity of Tea4-GFP (n > 130) measured at the cellular tips. (B) Extracts from cells 360 producing Rho1-HA (pREP4X-Rho1-HA) in the WT, rgf1Δ, and rgf1ΔPH cells were pulled down 361 with GST-C21RBD and blotted against anti-HA antibody (Rho1-GTP). Total Rho1-HA was 362 visualized by western blot (WCE). The relative units indicate the fold-differences in Rho1 levels .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 15 363 in the mutants compared with the WT strain, with an assigned value of 1 (bottom) from three 364 independent experiments. (C) Extracts from cells producing Rgf1-GFP or Rgf1ΔPH-GFP were 365 pulled down with GST-Tea4 purified from E. coli bound to GS-beads or with GS-beads alone and 366 blotted against anti-GST or anti-GFP antibodies (Pull-down). The total Rgf1 levels (WCE) were 367 visualized by western blot; tubulin was used as a loading control. (D) Extracts from cells 368 producing Rgf1-HA or Rgf1ΔPTTR-HA were pulled down with GST-Tea4 purified from E. coli 369 bound to GS-beads or with GS-beads alone and blotted against anti-GST or anti-GFP antibodies 370 (Pull-down). Total Rgf1 levels (WCE) were visualized by western blot; tubulin was used as a 371 loading control. (E) The percentage of cells WT, tea1Δ, rgf1Δ, rgf1ΔPH and rgf1ΔPTTR cells 372 forming branches 3 hours after release to growth after 3 days in stationary phase, in the absence 373 and in the presence of MBC (50 µg/ml). The mean ± SD of > 200 cells from two independent 374 experiments is shown. Statistical significance was calculated using a two-tailed unpaired 375 Student’s t test. ****P< 0.0001; **P < 0.01; *P < 0.05. 376 377 We wondered whether the Rgf1-ΔPH and Rgf1-ΔPTTR mutant proteins retained the 378 ability to bind Tea4 in vitro. Both proteins proficiently bound Tea4 in an in vitro pull-down assay 379 (Fig 4C and D). In vivo, we analyzed the percentage of T-shaped cells after re-entry from the 380 stationary phase to fresh medium. With MBC treatment, the percentage of T-shaped cells was 381 approximately 50% for the rgf1∆ mutant; this percentage was similar for the rgf1-ΔPH (~55%) 382 and rgf1-ΔPTTR (~35%) mutants and much higher than for the wild-type strain (~5%) (Fig 4E). 383 Taken together, these results indicate that the localization of Rgf1 to the PM and its ability to 384 activate Rho1 are closely linked: Both are required to maintain Tea4 at the cell tips and to 385 preserve the growth pattern after refeeding. Given that the Rgf1-ΔPTTR protein localized at the 386 growing end, where it is incompetent for Rho1 activation, GEF activity appears to be more 387 critical than GEF localization regarding Tea4 maintenance at the poles. Thus, Rho1 could act by .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 16 388 promoting the formation of a stable critical mass of Tea4 in the cell cortex that is essential to 389 activate pole-restricted growth. 390 391 Rgf1 is part of actin-dependent machinery that signals growth poles in the absence of the 392 Tea1–Tea4 complex 393 As we have described above, tea1Δ cells experience an exacerbated loss of polarity when 394 subjected to nutritional stress while keeping their cylindrical shape in normal conditions. We 395 wondered why this occurs and how cells recognize their poles in the absence of the classical Tea 396 markers. Since Tea4 localization is entirely dependent on Tea1(9) we utilized tea1Δ mutant to 397 eliminate both markers at the poles. It has been proposed that the high number of T-shaped 398 cells observed in the tea1Δ mutant in refeeding experiments (Fig 2B and 4E) is due to a transient 399 depolarization of the actin cytoskeleton (46). To address this issue, we confirmed the 400 disorganization of the actin cytoskeleton in wild-type cells grown stress treatments exposure to 401 KCl, sorbitol or heat, which promoted actin depolarization also induced branching in the tea1Δ 402 mutant for 3 days in liquid medium ( Fig 5A). In addition, ( Fig S5A and B) (11,47). We thought 403 that if the mechanism that keeps the identity of the growth sites in the absence of Tea1 is lost 404 because of transient actin depolarization, then treating tea1Δ cells with latrunculin A (LatA), 405 which prevents the polymerization of filamentous actin, should increase the number of 406 branched cells. As expected, 75% of the tea1Δ cells showed polarity defects during recovery 407 from the LatA treatment, compared with ~2% of the wild-type and untreated tea1Δ cells (Fig 408 5B). Thus, a properly polarized actin cytoskeleton is required to position growth sites at opposite 409 cell poles in the absence of Tea1–Tea4 markers. 410 411 Fig 5: Rgf1 is part of an actin-dependent machinery that signals growth poles in the absence 412 of the Tea1–Tea4 complex .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 17 413 (A) Images of LifeAct-GFP (actin) localization in the WT cells growing in liquid medium in the log- 414 phase or after 3 days in the stationary phase (s-phase). The maximum-intensity projection of six 415 Z-slides (0.5 µm step-size) of fluorescence is shown. (B) Morphology and quantitation of the T- 416 shaped wild-type and tea1Δ cells treated for 2 hours with DMSO (untreated) or 50 µM of 417 Latrunculin A (LatA) and then washed to allow growth for 3 hours. The graph represents the 418 mean ± SD of > 200 cells from three independent experiments. (C) Morphology and quantitation 419 of the T-shaped cells in the WT, tea1Δ, tea1Δ mod5Δ, and tea1Δ rgf1Δ cells grown to log phase 420 in YES liquid medium at 28°C. The graph represents the the mean ± SD of > 500 cells from two 421 independent experiments. (D) Representative images of Tea1-GFP, Tea4-GFP, GFP-Mod5 and 422 Rgf1-GFP localization in wild-type cells in the stationary phase after 3 days of growth in liquid 423 medium. The maximum-intensity projection of six Z-slides (0.5 µm step-size) of fluorescence is 424 shown. (E) Rgf1-GFP localization in WT cells growing in liquid medium untreated or treated with 425 KCl 0.6M, sorbitol 1.2M or 37°C (heat) for 1 hour. The maximum-intensity projection of six Z- 426 slides (0.5 µm step-size) of fluorescence is shown. The arrowheads point lateral accumulation of 427 Rgf1-GFP. (F) LifeAct-mCherry (actin in red) and Rgf1-GFP localization (green) in tea1Δ cells 428 treated with KCl 0.6M for 1 hour and then washed and allowed to grow without stress for the 429 indicated times. The maximum-intensity projection of four Z-slides (0.6 µm step-size) of 430 fluorescence is shown. The arrowheads point to lateral accumulation of Rgf1-GFP and actin (LA). 431 (G) Morphology of the tea1Δ and gef1Δtea1Δ cells growing in YES liquid medium. (H) 432 Quantitation of the T-shaped cells in WT and rgf1Δ cells treated with DMSO (-) or with MBC (50 433 µg/mL) for 4 hours. The graph represents the mean ± SD of > 200 cells from three independent 434 experiments. (I) Morphology and quantitation of the T-shaped cells in wild-type cells treated 435 with DMSO (WT -), MBC 50 µg/mL (WT MBC) for 4 hours, 50 µM of LatA for 2 hours and then 436 washed and allowed to grow with DMSO for 4 hours (WT LatA) or first treated with LatA, washed, 437 and then treated with 50 µg/mL of MBC for 4 hours (WT LatA-MBC). The graph represents the 438 mean ± SD of > 200 cells from three independent experiments. Statistical significance was .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 18 439 calculated using a two-tailed unpaired Student’s t test. ****P < 0.0001; ns = non-significant. 440 Scale bar 2 µm. 441 442 Given that Rgf1 is required for actin re-organization during NETO (28), we reasoned that 443 the absence of Rgf1 in the tea1Δ background could uncover polarity defects that would 444 otherwise remain undetected. This was indeed the case; 50% of the tea1Δ rgf1Δ cells in 445 unperturbed conditions were T-shaped compared with fewer than 5% of the tea1Δ and 446 tea1Δmod5Δ cells (Fig 5C). Thus, in the absence of Rgf1 and polarity markers (Tea1–Tea4), 447 stresses that induce actin depolarization were not necessary for branching to occur. Moreover, 448 Rho1 activation was required to maintain polarity, evidenced by the high number of T-shaped 449 cells observed in the tea1Δrgf1ΔPTTR mutant (Fig S5C). Therefore, in the absence of Tea1, Rgf1– 450 Rho1 probably mark the poles through an actin-dependent mechanism. 451 To understand why the tea1Δrgf1Δ null cells behaved alike tea1Δ stressed cell, we 452 studied the localization of Rgf1 under situations that depolarize the actin cytoskeleton. Rgf1- 453 GFP localization to the cell tips was lost quickly in cells treated with LatA, but was unaffected in 454 cells treated with MBC (Fig S5D). Thus, we reasoned that because actin is depolarized in 455 quiescent cells, Rgf1-GFP should behave similarly. Accordingly, Rgf1 was missing from the cell 456 periphery in stationary phase cells, while Tea1, Tea4, and Mod5 (which depend on MTs to reach 457 the poles), remained polarized in the same experiment (Fig 5D) (10,11,16). Moreover, Rgf1 458 disappeared from the cell tips under osmotic or heat stress but was observed in lateral patches 459 (Fig 5E). Then, we followed actin reorganization and Rgf1 localization during recovery from 460 osmotic stress in wild-type and tea1Δ cells. After relieving the stress, wild-type cells quickly re- 461 concentrated both actin and Rgf1 at the poles ( Fig S5E and Movie S3). However, in tea1Δ cells 462 actin and Rgf1 frequently localized at the lateral cortex precisely at points where a branch began 463 to grow (Fig 5F and Movie S4). .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 19 464 To find out whether other proteins of the growth machinery also define the growth sites 465 or if it is a specific characteristic of Rgf1, we used a mutant lacking gef1. Gef1 is a GEF of Cdc42 466 involved in polarized growth and undergoes a similar relocation from poles to lateral patches 467 under stress, where it is required to activate Cdc42 (48,49). Interestingly, the gef1Δtea1Δ double 468 mutant displayed a similar percentage of T-shaped cells as the tea1Δ mutant, indicating that in 469 the absence of Tea1, Gef1 is not necessary for marking the poles, in contrast to Rgf1–Rho1 (Fig 470 5G and S5F). 471 Taken together, these results indicated that Rgf1 localization at the cell tips depends on a 472 properly polarize actin cytoskeleton but it is independent on MTs. Thus, both actin 473 concentration and Rho1 activation at the poles are necessary events to define these locations 474 as growth sites. 475 476 Tea1–Tea4-MTs and Rgf1-Rho-actin define two parallel pathways to restrict growth to the cell 477 tips 478 Because stresses that induce actin disorganization and Rgf1 delocalization at the poles increase 479 the percentage of T-shaped cells in the tea1Δ mutant (Fig 5D and S5A), we wondered whether 480 chemical ablation of the pathway that delivers Tea1 and Tea4 to the cell tips would yield a similar 481 result. When we treated the rgf1Δ mutant with MBC, ~30% of the cells exhibited a branched 482 phenotype compared with ~5% of the wild-type cells (Fig 5H and S5G). Therefore, in the absence 483 of Rgf1 (which lacks the “actin-dependent signal” necessary to recognize the polarity growth 484 zones) and MTs, the imposition of stress is not necessary for branching to occur. The previous 485 results suggest that there could be two parallel pathways for positioning the growth poles: one 486 dependent on MTs and the Tea1–Tea4 markers, and another dependent on actin and Rgf1– 487 Rho1. To mimic the elimination of both signaling pathways by chemical treatment, we treated 488 wild-type cells in the log phase with LatA for 2 hours (to block actin-dependent signaling), 489 washed them, and exposed them to MBC for 4 hours (to remove MTs) to analyze the number of .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 20 490 T-shaped cells. Only after the treatment with both, LatA and MBC, ~35% of the cells showed 491 branches (Fig 5I). 492 Our results indicate that Rgf1–Rho1 and Tea1–Tea4 are part of the same complex, with 493 similar functions to delimit the sites of polarized growth of S. pombe. We propose that Rgf1 and 494 Rho1 could activate an actin-dependent pathway that instructs cells to grow at the tips 495 regardless of the classical polarity markers. 496 497 498 DISCUSSION 499 Cross-talk between microtubules (MT) and the actin cytoskeleton is crucial for various cellular 500 processes, including asymmetric cell division, the establishment of cell growth zones, and cell 501 migration. In fission yeast, MTs deliver the Tea1–Tea4 complex to the cell tips, where actin 502 concentrates to promote growth. Tea1–Tea4 act as end markers, defining the organization of 503 cell-growth zones and, consequently, the direction of growth. While it is known that these 504 polarity markers are not essential for organizing a growth zone, they become critical for placing 505 the growth zone correctly, especially under stress conditions. However, certain questions 506 remain unanswered, such as how and why the Tea1–Tea4 complex remains linked to the plasma 507 membrane (PM) once the growth direction is established and how cells mark their tips in the 508 absence of Tea1. In the present study, we addressed some of these questions by demonstrating 509 that Rgf1 functions as a molecular link between the Tea1-Tea4 complex and the PM. Through 510 Rho1 activation, Rgf1 stabilizes Tea4 at the cell ends, promoting its accumulation. Additionally, 511 we described an alternative actin-dependent mechanism, driven by Rgf1 and Rho1, for marking 512 the poles independently to the known MT- and Tea-dependent pathway. 513 514 Rgf1 (Rho1 GEF) activity toward Rho1 is required for stable accumulation of Tea4 at the cell 515 ends .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 21 516 Failure to accumulate Tea4 at the cell cortex in the rgf1Δ cells is independent of the movement 517 of Tea4 riding on the tips of polymerizing MTs. Wild-type and rgf1Δ cells exhibited similar rates 518 of movement for Tea4-GFP dots from the middle of the cell to the cell ends (Fig S1D). However, 519 once the Tea4-GFP dots had reached the cell ends, their fluorescence faded in the rgf1Δ cells. 520 The refill mechanism responsible for keeping Tea4 stacked at the growing pole depends on 521 correct binding of Rgf1 to the PM and on Rgf1 catalytic activity toward Rho1. We provided 522 evidence that Rgf1 physically binds the polarity marker complex through its interaction with 523 Tea4 (Fig 3B-D). Moreover, Rgf1 binds to PM phospholipids, likely through its PH domain 524 interacting with membrane PI4P. The physical association between Rgf1 and Tea4 is crucial for 525 the localization of Tea4 and Tea1 at the cell poles. Indeed, a mutant of Rgf1 lacking the 526 membrane-binding domain, which binds Tea4 in vitro, exhibited similar polarity defects as the 527 null mutant (Fig 4E). In the rgf1ΔPH mutant, the reduced activation of Rho1 (Fig 4B) is coupled 528 to protein instability (Fig 3H) (probably because it is not properly bound to the PM) which makes 529 the interpretation of this result challenging. Additionally, a catalytic mutant (rgf1-ΔPTTR) also 530 shows impaired Tea4 anchoring. The Rgf1-ΔPTTR protein binds Tea4 in vitro, localizes to the 531 growing end and is catalytically deficient (36), indicating that Rho1 activity is required to 532 maintain a the Tea1-4 complex stable at the pole. Furthermore, the tandem Rgf1–Rho1 also 533 affects Tea1 functions, promoting MT catastrophe once they reach the cell pole or restricting 534 growth sites after refeeding (Figs 1G-H and 2B). To maintain polarity markers stably at the poles, 535 Rgf1 functions together with Mod5. The loss of Tea4 localization at the poles is partial in the 536 rgf1Δ and mod5Δ single mutants and complete when both mutations are combined, causing a 537 defect in polarity even in cells with a continuous supply of Tea1-Tea4 (Fig 2B and C) . It is likely 538 that Mod5 and Rgf1 anchor the Tea1-4 complex to the PM through different proteins: Mod5 539 interacts with Tea1 (17), while Rgf1 interacts with Tea4. This mechanism ensures double 540 anchoring of the Tea1–Tea4 complex to the PM; thus, when one of these connections is lost, the .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 22 541 other remains available. Only when both are missing do the polarity markers completely lose 542 their connection with the PM, mimicking the behavior of a the tea1Δ mutant. 543 544 Rgf1–Rho1 are involved in defining the growth sites 545 Another question that must be addressed is how cells detect their growth sites in the absence 546 of Tea markers. Here we proposed that actin and Rho1 activation by Rgf1 are behind this 547 process. We observed two ways to induce the formation of branched cells in the tea1Δ 548 background: one in the presence of certain stresses, “the stress pathway” and the other in the 549 absence of rgf1, “Rgf1 depletion” (Fig 5C and S5B). The loss of polarity induced by “the stress 550 pathway” also leads to disorganization of the actin cytoskeleton and, consequently, Rgf1 551 delocalization. Therefore, both pathways lead to insufficient activation of Rho1 at the poles. 552 When the tea1Δ cells recover from stress, Rgf1 and actin appear simultaneously at sites 553 were ectopic growth occurs. The interdependence of actin and Rgf1 localization (Fig S5D) (28) 554 suggests a positive feedback loop between Rho1 activation and actin organization at the growth 555 sites. Therefore, Rgf1–Rho1 would act as Tea markers, defining growth sites without directly 556 promoting growth, given that the tea1Δrgf1Δ double mutant retains the ability to grow, albeit 557 at the wrong places. Interestingly, in the absence of Tea1 and Gef1 (Cdc42 GEF), which is also 558 involved in polarized growth and relies on actin for proper localization (48–50), cells do not form 559 branches, suggesting that is the activation of Rho1 and not that of Cdc42 which specifically 560 defines the growth sites. 561 The interaction between Tea4 and Rgf1–Rho1 indicates their involvement in the same 562 complex, with each protein essential for the accumulation of the other at the poles. Rgf1 563 together with Mod5 acts to link Tea4 to the membrane, and Tea1-4 assists in returning Rgf1 to 564 the poles after stress-induced actin disorganization (Fig 5F and S5E ), establishing a functional 565 link between MTs and actin cytoskeletons (Fig 6A). 566 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 23 567 Fig 6: Rgf1-Rho1 functions as a molecular link between Tea4 and the PM and marks the growth 568 sites in an actin-dependent manner 569 (A) In wild-type cells MT deliver Tea1 and Tea4 to the cell poles, where they bind to the 570 membrane through their interaction with Mod5 and Rgf1, respectively. Rgf1, in turn, interacts 571 with the PM due to its affinity for the phospholipid PI4P and activates Rho1, promoting proper 572 actin cytoskeleton polarization at the cellular tips. (B) Cells lacking Tea1 still recognize the cell 573 poles correctly because Rgf1 is localized in these regions, where it activates Rho1, thus allowing 574 the maintenance of polarized actin. (C) In rgf1Δ cells Tea1-Tea4 partially disappears from the 575 pole, although a remnant remains attached to Mod5. Rho1 would be inactive, leading to actin 576 cytoskeleton disorganization. However, the continuous supply of Tea1–Tea4 from MTs persists, 577 allowing the cells to grow correctly. (D) When cells lack tea1 and rgf1, they lose both pathways 578 that allowed them to distinguish the tips. First, there is no continuous supply of polarity markers 579 towards the cell ends by the MTs that mark the growth sites. Second, Rgf1 is not at the pole to 580 activate Rho1, leading to actin disorganization. The elimination of both pathways causes the 581 cells to direct their growth towards incorrect locations, where growth factors probably 582 accumulate. 583 584 We showed that Rgf1 and Rho1 are required to maintain an actin-dependent signal that 585 preserves the identity of the cell poles, which becomes evident in the absence of the Tea 586 markers. This would explain why the rgf1Δtea1Δ double mutant forms T-shaped cells in the log 587 phase, without stresses or refeeding treatments. We propose that there are two different 588 pathways to choose the growth sites under different environmental conditions: the canonical 589 pathway dependent on MT and Tea1–Tea4 and, a novel pathway dependent on actin and Rgf1– 590 Rho1. Various combinations of mutants and/or chemical ablation of one component from each 591 pathway (Tea1–Tea4–MT and Rgf1–Rho1–actin) at once, leads to comparable outcomes. For 592 example, chemical actin depolymerization in the tea1Δ cells induces branching (Fig 5B), whereas .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 24 593 treatment of the rgf1Δ cells with MBC to prevent the constant supply of Tea markers generates 594 T-shaped cells as well (Fig 5H). Furthermore, wild-type cells treated to transiently depolarize 595 actin and subsequently prevented for refeeding of polarity markers to the tips (via MT de- 596 polymerization) also experience difficulties in detecting the cell poles (Fig 5I). Thus, cells would 597 have different pathways to maintain their cylindrical morphology even if one of them is 598 challenged by internal or external conditions. When the continuous supply of Tea1 from MT is 599 disrupted (Fig 6B), cells could use an additional cue provided for a different cytoskeletal polymer, 600 actin. Similarly, when actin becomes disorganized, for example, in the transition from 601 monopolar to bipolar growth (NETO) or in rgf1Δ mutant (defective in NETO), cells would count 602 on polarity markers transported by MTs (Fig 6C). Only when both, the actin and MT 603 cytoskeletons are compromised, fission yeast cells lose their cylindrical shape (Fig 6D). This 604 sophisticated regulation highlights the importance of maintaining cell morphology throughout 605 the cell cycle and under changing environmental conditions. 606 607 608 MATERIALS AND METHODS 609 Media, Reagents and Genetics 610 S. pombe strains were streaked on plates of complete yeast growth medium (YES), or selective 611 medium (EMM) supplemented with the appropriate requirements (51), and incubated at 28°C 612 until colonies formed. For each biological replicate, a single colony was used to inoculate 5 mL 613 of the respective liquid media. Cultures were incubated at 28°C overnight with shaking (200 614 rpm). Each overnight culture was subsequently used as a seed culture to inoculate fresh media. 615 Fresh cultures were next grown at 28°C, 200 rpm, to OD 600 = 0.5-0.6 at the time of harvest. 616 Crosses were performed by mixing the appropriate strains directly on sporulation medium 617 plates. Recombinant strains were obtained by tetrad analysis or the “random spore” method. 618 For overexpression experiments using the nmt1 promoter, cells were grown in EMM containing .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 25 619 15 µM thiamine up to the logarithmic phase. Then, the cells were harvested, washed three times 620 with water, and inoculated in fresh medium (without thiamine) at an optical density at 600 nm 621 (OD600) of 0.01 for 18-20 hours. Two-hybrid interaction was tested with YNB medium lacking 622 histidine in Saccharomyces cerevisiae strain AH109 (Clontech). 623 624 Plasmid and DNA manipulations 625 Plasmids used in this study are listed in Key resources table (Recombinant DNA). pREP4x-HArho1 626 (with thiamine-repressible nmt1 promoter) and pGEX-C21RBD plasmids (rhotekin-binding 627 domain) kindly provided by Pilar Pérez (Instituto de Biología Funcional y Genómica, Salamanca, 628 Spain) were used to detect Rho1-GTP levels. To express proteins in Escherichia coli we used a 629 pGEX-2T plasmid that contains a GST to tagged genes al 5’. pGEX-rgf1 (pGR152) and pGEX-tea4 630 (pGR129) were made by inserting the entire ORF of rgf1 (without introns) or tea4 in frame into 631 pGEX-2T and purified for be used in pull-down assays. To perform lipid strips assays we 632 constructed the plasmids pGR128 that contains the ORF of rgf1 (from aminoacid 1-922) without 633 the last 428 aminoacids containing the CNH domain, and pGR138 that contains the ORF of rgf1 634 (from aminoacid 1-722) without the last 578 aminoacids containing the PH and CNH domains. 635 For two-hybrid experiments we constructed the plasmids pGAD-tea1 (pGR135), pGAD-tea4 636 (pGR106), and pGBK-rgf1 (pRZ97) where the entire ORF of the corresponding gene was inserted 637 in frame into the pGADT7 (GAL4 activation domain) or pGBKT7 (GAL4 binding domain) plasmid 638 (Clontech). 639 640 Protein extracts and immunoblot analysis 641 S. pombe cultures (5 mL) at an OD600 of 0.5 were pelleted just after the addition of 10% TCA and 642 washed in 20% TCA. The pellets were resuspended in 100 μl 12.5% TCA with the addition of glass 643 beads and lysed by vortexing for 5 min. Cell lysates were pelleted, washed in iced acetone, and 644 dried at 55°C for 15 min. Pellets were resuspended in 50 μl of a solution containing 1% SDS, 100 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 26 645 mM Tris–HCl (pH 8.0), and 1 mM EDTA. Samples were electrophoretically separated by SDS- 646 PAGE (4–15% MiniProtein Gel, BioRad) and immunodetected with anti-GFP (Living Colors, 647 RRID:AB_10013427) and anti-mouse (Bio-Rad, AB_11125547) antibodies. As a loading control, 648 we used monoclonal antitubulin antibodies (Sigma, RRID:AB_477579). 649 650 Immunoprecipitations and pull-down assays 651 Immunoprecipitation assays were performed as described previously with some modifications 652 (52). Briefly, logarithmic cultures were pelleted and re-suspended in lysis buffer (10 mM Tris– 653 HCl pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 0.5% NP40, containing 100 μM PMSF, leupeptin, and 654 aprotinin) and lysed in a cryogenic grinder. Lysates were centrifuged for 5 min at 6000 g, and 655 then 10 μL of GFP-Trap magnetic beads (Chromotek) was added to the supernatants an 656 incubated for 1 hour at 4°C. Immunoprecipitates were washed three times with dilution buffer 657 (10 mM Tris–HCl pH 7.5, 150 mM NaCl, and 0.5 mM EDTA). Proteins were released from 658 immunocomplexes by boiling for 5 minutes in sodium dodecyl sulfate (SDS) loading buffer. 659 Samples were separated by SDS–polyacrylamide gel electrophoresis (4–15% Mini-Protean TGX 660 gels, Bio-Rad) and detected by immunoblotting with polyclonal anti-HA (Roche, 661 RRID:AB_514506) or anti-GFP antiserum (Living Colors, RRID:AB_10013427). For pull-down 662 assays, first GST-tagged Rgf1, Tea4, or C21RBD (rhotekin-binding domain to detect Rho1-GTP 663 levels) was purified from E. coli (BL21). The fusion proteins were produced by adding 0.5 mM 664 IPTG at 18°C overnight (or 3 hours at 28°C for C21RBD). Cells were sonicated and proteins 665 immobilized on glutathione-Sepharose (GS) 4B beads (GE-Healthcare). After incubation for 1 666 hour, the beads were washed several times, and the bound proteins were analyzed by SDS-PAGE 667 and stained with Coomassie brilliant blue. Pull-down assays were performed as described 668 previously (53). In brief, extracts from the indicated strains were obtained by using 500 μL of 669 lysis buffer (50 mM Tris–HCl pH 7.5, 20 mM NaCl, 0.5% NP-40, 10% glycerol, 0.1 mM 670 dithiothreitol, and 2 mM Cl2Mg, containing 100 μM PMSF, leupeptin, and aprotinin) and lysed in .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 27 671 a cryogenic grinder. Cell extracts (2–3 mg of total protein) were incubated with 2–10 μg of GST- 672 tagged protein coupled to GS beads for 2 hours, washed four times with lysis buffer, and blotted 673 with an anti-HA or anti-GFP antibody. Protein levels in whole-cell extracts (80 μg of total protein) 674 were monitored by western blot. Tubulin and GST were used as loading controls. 675 676 Lipid strip overlay assays 677 Lipid strip overlay assays were performed as described previously (54) using lipid strip 678 membranes (p-6002, Echelon). Strips were blocked with 3% fatty acid–free bovine serum 679 albumin (BSA; Sigma) in TBS-T (10 mM Tris–HCl pH 8.0, 150 mM NaCl, and 0.1% Tween-20; TBST- 680 BSA) at room temperature for 1 hour and then incubated for 2 hours with 1 μg/mL GST, GST- 681 Rgf1, or GST-Rgf1ΔPH in TBST-BSA. The strips were then washed three times with 5 mL of TBST- 682 BSA and incubated with anti-GST horseradish peroxidase–conjugated antibody (GE Healthcare, 683 RRID:AB_771429) diluted in TBST-BSA. Bound protein was detected using an enhanced 684 chemoluminescence detection kit (BioRad). GST, GST-Rgf1, and GST-Rgf1ΔPH were expressed in 685 E. coli (BL21) and purified with GS beads (GE-Healthcare) according to the manufacturer’s 686 instructions as described above. Once attached to the beads, they were washed and eluted with 687 200 μL of elution buffer (100 mM Tris–HCl pH 8.0, 120 mM NaCl) with 20 mM of L-glutathione 688 reduced (Sigma) freshly added for 30 minutes at 4°C. Aliquots were frozen in liquid nitrogen with 689 15% of glycerol and stored at -80°C until use. 690 691 Microscopy 692 Wet preparations were observed with an Andor Dragonfly 200 Spinning-disk confocal 693 microscope equipped with a sCMOS Sona 4.2B-11 camera (Andor) and controlled with the 694 Fusion 2.2 acquisition software; or a Personal Deltavision (Applied Precision, LLC) microscope 695 equipped with a CoolSNAP HQ2 camera (Photometrics) and controlled with softWoRx Resolve 696 3D. Depending on the experiment, a single focal plane at the centre of the cell or a stack of 4–6 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 28 697 images covering the entire volume of the cell (Z-series) with a spacing of 0.5–0.6 μm were 698 captured, and the maximum projection was generated. 699 To analyze protein dynamics, time-lapse experiments were performed with cells in μ- 700 Slide 8-well (Ibidi) coated with soybean lectin (Sigma Aldrich) and imaged at the indicated times 701 using an Andor Dragonfly Spinning-disk confocal microscope. We used the ImageJ 1.53t software 702 to calculate the relative fluorescence intensity of each protein at the cell tips. We have designed 703 an ImageJ macro to automatically select the fluorescence regions (ROI manager) of the cell poles 704 and to measure the fluorescence intensity of 40–130 cells. We used the integrated density value 705 of each tip of the different strains versus the average of the integrated density of the wild-type 706 strain to calculated the relative fluorescence levels. To create kymographs, we drew a line from 707 the cell’s centre to the pole along a microtubule on a time-lapse movie and utilized the 708 KymographBuilder plugin of the ImageJ software. For super-resolution radial fluctuations (SRRF) 709 images of Tea4-GFP (green) and mCherry-Atb2 (red) in “head-on” cell tips, cells were mounted 710 onto a μ-Slide pre-coated with lectin. Bound cells found to be frontally arrayed (“head-on”) were 711 visually selected for imaging using a the SRRF module of the Andor Dragonfly Confocal 712 microscope. Fifty repetitions of one focal plane were taken for each time point. The time 713 projection of the three images at different time points is shown to follow Tea4 cluster 714 movement. 715 Calcofluor white (Blankophor BBH, Bayer Corporation) staining was performed by 716 adding 1 μL of a stock solution (2.5 mg/mL) to 500 μL of samples for 20 seconds, followed by a 717 wash with phosphate–buffered saline (PBS). 718 719 Quantification and statistical analysis 720 Statistical analyses and graphs were generated using GraphPad Prism Software version 9.5.1. To 721 compare two conditions, a two-tailed unpaired Student’s t-test was applied to determine 722 statistical significance (as detailed in the Fig legends). P < 0.05 were considered significant. The .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 29 723 graphs show the mean ± standard deviation of the indicated data. Asterisks represent the 724 following: *P < 0.05 **P < 0.01; ***P < 0.001 ****P < 0.0001. 725 726 Author contributions 727 Conceptualization, P.G., and Y.S.; Methodology, P.G.; Investigation, P.G., R.C. and T.E.; Writing – 728 Original Draft, P.G.; Writing – Review & Editing, P.G., and Y.S.; Funding Acquisition, Y.S.; 729 Supervision, P.G. and Y.S. 730 731 Acknowledgments 732 We thank J.C. Ribas, Phong T. Tran, Sophie Martin, Sergio Rincón, Ken Sawin, Paul Nurse, James 733 Moseley, Kathleen L. Gould, César Roncero, Henar Valdivieso, Sergio Moreno and Pilar Pérez for 734 sharing strains and plasmids. We also wish to thank Javier Encinar del Dedo and Sergio Rincón 735 for their very helpful comments on the manuscript. We are grateful to Jesús Pinto (IBFG 736 bioinformatics facility) for ImageJ macro used for fluorescence quantification. 737 738 739 REFERENCES 740 1. Allam AH, Charnley M, Russell SM. Context-Specific Mechanisms of Cell Polarity 741 Regulation. J Mol Biol. 2018 Sep;430(19). 742 2. Drubin DG, Nelson WJ. Origins of cell polarity. Cell. 1996;84:335–44. 743 3. Hoffman CS, Wood V, Fantes PA. An Ancient Yeast for Young Geneticists: A Primer on 744 the Schizosaccharomyces pombe Model System. Genetics [Internet]. 2015 745 Oct;201(2):403 LP-- 423. Available from: 746 http://www.genetics.org/content/201/2/403.abstract .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 30 747 4. Chang F, Martin SG. Shaping fission yeast with microtubules. Cold Spring Harb Perspect 748 Biol [Internet]. 2009;1(1):a001347. Available from: 749 https://www.ncbi.nlm.nih.gov/pubmed/20066076 750 5. Huisman SM, Brunner D. Cell polarity in fission yeast: a matter of confining, positioning, 751 and switching growth zones. Semin Cell Dev Biol [Internet]. 2011;22(8):799–805. 752 Available from: https://www.ncbi.nlm.nih.gov/pubmed/21803169 753 6. Martin SG. Microtubule-dependent cell morphogenesis in the fission yeast. Trends Cell 754 Biol. 2009;19:447–54. 755 7. Mitchison JM, Nurse P. Growth in cell length in the fission yeast Schizosaccharomyces 756 pombe. J Cell Sci [Internet]. 1985;75:357–76. Available from: 757 https://www.ncbi.nlm.nih.gov/pubmed/4044680 758 8. Bähler J, Pringle JR. Pom1p, a fission yeast protein kinase that provides positional 759 information for both polarized growth and cytokinesis. Genes Dev. 1998;12:1356–70. 760 9. Martin SG, McDonald WH, Yates 3rd JR, Chang F. Tea4p links microtubule plus ends 761 with the formin for3p in the establishment of cell polarity. Dev Cell [Internet]. 762 2005;8(4):479–91. Available from: https://www.ncbi.nlm.nih.gov/pubmed/15809031 763 10. Mata J, Nurse P. tea1 and the microtubular cytoskeleton are important for generating 764 global spatial order within the fission yeast cell. Cell [Internet]. 1997;89(6):939–49. 765 Available from: https://www.ncbi.nlm.nih.gov/pubmed/9200612 766 11. Tatebe H, Shimada K, Uzawa S, Morigasaki S, Shiozaki K. Wsh3/Tea4 is a novel cell-end 767 factor essential for bipolar distribution of Tea1 and protects cell polarity under 768 environmental stress in S. pombe. Curr Biol [Internet]. 2005;15(11):1006–15. Available 769 from: https://www.ncbi.nlm.nih.gov/pubmed/15936270 770 12. Browning H, Hayles J, Mata J, Aveline L, Nurse P, McIntosh JR. Tea2p is a kinesin-like 771 protein required to generate polarized growth in fission yeast. J Cell Biol. 2000;151:15– 772 27. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 31 773 13. Browning H, Hackney DD, Nurse P. Targeted movement of cell end factors in fission 774 yeast. Nat Cell Biol [Internet]. 2003/08/02. 2003;5(9):812–8. Available from: 775 http://www.ncbi.nlm.nih.gov/pubmed/12894167 776 14. Brunner D, Nurse P. CLIP170-like tip1p spatially organizes microtubular dynamics in 777 fission yeast. Cell. 2000;102:695–704. 778 15. Busch KE, Hayles J, Nurse P, Brunner D. Tea2p kinesin is involved in spatial microtubule 779 organization by transporting tip1p on microtubules. Dev Cell. 2004;6:831–43. 780 16. Snaith H, Sawin KE. Fission yeast mod5p regulates polarized growth through anchoring 781 of tea1 at the cell tips. Nature. 2003;423:647–51. 782 17. Snaith HA, Samejima I, Sawin KE. Multistep and multimode cortical anchoring of tea1p 783 at cell tips in fission yeast. EMBO J [Internet]. 2005/10/14. 2005;24(21):3690–9. 784 Available from: http://www.ncbi.nlm.nih.gov/pubmed/16222337 785 18. Dodgson J, Chessel A, Yamamoto M, Vaggi F, Cox S, Rosten E, et al. Spatial segregation 786 of polarity factors into distinct cortical clusters is required for cell polarity control. Nat 787 Commun [Internet]. 2013/05/16. 2013;4:1834. Available from: 788 http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Cit 789 ation&list_uids=23673619 790 19. Feierbach B, F. V, Chang F. Regulation of a formin complex by the microtubule plus end 791 protein tea1p. J Cell Biol. 2004;165(697–707). 792 20. Bi E, Park HO. Cell polarization and cytokinesis in budding yeast. Genetics [Internet]. 793 2012;191(2):347–87. Available from: http://www.ncbi.nlm.nih.gov/pubmed/22701052 794 21. Hodge RG, Ridley AJ. Regulating Rho GTPases and their regulators. Nat Rev Mol Cell Biol 795 [Internet]. 2016;17(8):496–510. Available from: 796 https://www.ncbi.nlm.nih.gov/pubmed/27301673 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 32 797 22. Martin SG, Arkowitz RA. Cell polarization in budding and fission yeasts. FEMS Microbiol 798 Rev [Internet]. 2014;38(2):228–53. Available from: 799 https://www.ncbi.nlm.nih.gov/pubmed/24354645 800 23. Machacek M, Hodgson L, Welch C, Elliott H, Pertz O, Nalbant P, et al. Coordination of 801 Rho GTPase activities during cell protrusion. Nature. 2009 Sep 19;461(7260). 802 24. Bendezu FO, Vincenzetti V, Vavylonis D, Wyss R, Vogel H, Martin SG. Spontaneous 803 Cdc42 polarization independent of GDI-mediated extraction and actin-based trafficking. 804 PLoS Biol [Internet]. 2015;13(4):e1002097. Available from: 805 http://www.ncbi.nlm.nih.gov/pubmed/25837586 806 25. Haupt A, Ershov D, Minc N. A Positive Feedback between Growth and Polarity Provides 807 Directional Persistency and Flexibility to the Process of Tip Growth. Current Biology. 808 2018 Oct;28(20). 809 26. Johnson JM, Jin M, Lew DJ. Symmetry breaking and the establishment of cell polarity in 810 budding yeast. Curr Opin Genet Dev [Internet]. 2011;21(6):740–6. Available from: 811 http://www.ncbi.nlm.nih.gov/pubmed/21955794 812 27. Arellano M, Duran A, Perez P. Localisation of the Schizosaccharomyces pombe rho1p 813 GTPase and its involvement in the organisation of the actin cytoskeleton. J Cell Sci. 814 1997;110(20):2547–55. 815 28. Garcia P, Tajadura V, Garcia I, Sanchez Y. Rgf1p is a specific Rho1-GEF that coordinates 816 cell polarization with cell wall biogenesis in fission yeast. Mol Biol Cell [Internet]. 817 2006;17(4):1620–31. Available from: https://www.ncbi.nlm.nih.gov/pubmed/16421249 818 29. Garcia P, Garcia I, Marcos F, de Garibay GR, Sanchez Y. Fission yeast rgf2p is a rho1p 819 guanine nucleotide exchange factor required for spore wall maturation and for the 820 maintenance of cell integrity in the absence of rgf1p. Genetics [Internet]. 821 2009;181(4):1321–34. Available from: 822 https://www.ncbi.nlm.nih.gov/pubmed/19189958 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 33 823 30. Davidson R, Laporte D, Wu JQ. Regulation of Rho-GEF Rgf3 by the arrestin Art1 in fission 824 yeast cytokinesis. Mol Biol Cell [Internet]. 2014/12/05. 2015;26(3):453–66. Available 825 from: https://www.ncbi.nlm.nih.gov/pubmed/25473118 826 31. Morrell-Falvey JL, Ren L, Feoktistova A, Haese GD, Gould KL. Cell wall remodeling at the 827 fission yeast cell division site requires the Rho-GEF Rgf3p. J Cell Sci [Internet]. 828 2005;118(Pt 23):5563–73. Available from: 829 https://www.ncbi.nlm.nih.gov/pubmed/16291723 830 32. Mutoh T, Nakano K, Mabuchi I. Rho1-GEFs Rgf1 and Rgf2 are involved in formation of 831 cell wall and septum, while Rgf3 is involved in cytokinesis in fission yeast. Genes Cells 832 [Internet]. 2005;10(12):1189–202. Available from: 833 https://www.ncbi.nlm.nih.gov/pubmed/16324155 834 33. Tajadura V, Garcia B, Garcia I, Garcia P, Sanchez Y. Schizosaccharomyces pombe Rgf3p 835 is a specific Rho1 GEF that regulates cell wall beta-glucan biosynthesis through the 836 GTPase Rho1p. J Cell Sci [Internet]. 2004;117(Pt 25):6163–74. Available from: 837 https://www.ncbi.nlm.nih.gov/pubmed/15546915 838 34. García P, Celador R, Pérez-Parrilla J, Sánchez Y. Fission Yeast Rho1p-GEFs: From Polarity 839 and Cell Wall Synthesis to Genome Stability. Int J Mol Sci. 2022 Nov 11;23(22). 840 35. Pérez P, Cortés JCG, Cansado J, Ribas JC. Fission yeast cell wall biosynthesis and cell 841 integrity signalling. Cell Surf [Internet]. 2018 Dec 1 [cited 2021 Jan 18];4:1–9. Available 842 from: http://www.ncbi.nlm.nih.gov/pubmed/32743131 843 36. Garcia P, Tajadura V, Sanchez Y. The rho1p exchange factor Rgf1p signals upstream 844 from the Pmk1 mitogen-activated protein kinase pathway in fission yeast. Mol Biol Cell. 845 2009;20(2). 846 37. Feierbach B, Chang F. Roles of the fission yeast formin For3 in cell polarity, actin cable 847 formation and symmetric cell division. Curr Biol. 2001;11:1656–65. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 34 848 38. Glynn JM, Lustig RJ, Berlin A, Chang F. Role of bud6p and tea1p in the interaction 849 between actin and microtubules for the establishment of cell polarity. Curr Biol. 850 2001;11:836–45. 851 39. Niccoli T, Arellano M, Nurse P. Role of Tea1p, Tea3p and Pom1p in the determination of 852 cell ends in Schizosaccharomyces pombe. Yeast [Internet]. 2003/12/10. 853 2003;20(16):1349–58. Available from: http://www.ncbi.nlm.nih.gov/pubmed/14663827 854 40. Bicho CC, Kelly DA, Snaith HA, Goryachev AB, Sawin KE. A catalytic role for Mod5 in the 855 formation of the Tea1 cell polarity landmark . Curr Biol. 2010;20:1752–7. 856 41. Bähler J, Steever AB, Wheatley S, Wang Y, Pringle JR, Gould KL, et al. Role of polo kinase 857 and Mid1p in determining the site of cell division in fission yeast. J Cell Biol. 858 1998;143:1603–16. 859 42. Yu JW, Mendrola JM, Audhya A, Singh S, Keleti D, DeWald DB, et al. Genome-Wide 860 Analysis of Membrane Targeting by S. cerevisiae Pleckstrin Homology Domains. Mol 861 Cell. 2004 Mar;13(5). 862 43. Singh N, Reyes-Ordoñez A, Compagnone MA, Moreno JF, Leslie BJ, Ha T, et al. 863 Redefining the specificity of phosphoinositide-binding by human PH domain-containing 864 proteins. Nat Commun. 2021 Jul 15;12(1). 865 44. Audhya A, Emr SD. Stt4 PI 4-Kinase Localizes to the Plasma Membrane and Functions in 866 the Pkc1-Mediated MAP Kinase Cascade. Dev Cell. 2002 May;2(5). 867 45. Snider CE, Willet AH, Chen JS, Arpağ G, Zanic M, Gould KL. Phosphoinositide-mediated 868 ring anchoring resists perpendicular forces to promote medial cytokinesis. Journal of 869 Cell Biology. 2017 Oct 2;216(10). 870 46. Sawin KE, Snaith HA. Role of microtubules and tea1p in establishment and maintenance 871 of fission yeast cell polarity. J Cell Sci [Internet]. 2004/01/22. 2004;117(Pt 5):689–700. 872 Available from: http://www.ncbi.nlm.nih.gov/pubmed/14734657 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 35 873 47. Robertson AM, Hagan IM. Stress-regulated kinase pathways in the recovery of tip 874 growth and microtubule dynamics following osmotic stress in S. pombe. J Cell Sci. 2008 875 Dec 15;121(24). 876 48. Salat-Canela C, Carmona M, Martín-García R, Pérez P, Ayté J, Hidalgo E. Stress- 877 dependent inhibition of polarized cell growth through unbalancing the GEF/GAP 878 regulation of Cdc42. Cell Rep. 2021 Nov;37(5). 879 49. Coll PM, Trillo Y, Ametzazurra A, Perez P. Gef1p, a new guanine nucleotide exchange 880 factor for Cdc42p, regulates polarity in Schizosaccharomyces pombe. Mol Biol Cell 881 [Internet]. 2003;14(1):313–23. Available from: 882 https://www.ncbi.nlm.nih.gov/pubmed/12529446 883 50. Hirota K, Tanaka K, Ohta K, Yamamoto M. Gef1p and Scd1p, the Two GDP-GTP 884 exchange factors for Cdc42p, form a ring structure that shrinks during cytokinesis in 885 Schizosaccharomyces pombe. Mol Biol Cell [Internet]. 2003;14(9):3617–27. Available 886 from: https://www.ncbi.nlm.nih.gov/pubmed/12972551 887 51. Moreno S, Klar A, Nurse P. [56] Molecular genetic analysis of fission yeast 888 Schizosaccharomyces pombe. In 1991. p. 795–823. Available from: 889 https://linkinghub.elsevier.com/retrieve/pii/007668799194059L 890 52. Calvo IA, García P, Ayté J, Hidalgo E. The transcription factors Pap1 and Prr1 collaborate 891 to activate antioxidant, but not drug tolerance, genes in response to H 2O 892 2. Nucleic Acids Res. 2012;40(11). 893 53. García P, Coll PM, del Rey F, Geli MI, Pérez P, Vázquez de Aldana CR, et al. Eng2, a new 894 player involved in feedback loop regulation of Cdc42 activity in fission yeast. Sci Rep. 895 2021;11(1). 896 54. Fernández-Golbano IM, Idrissi FZ, Giblin JP, Grosshans BL, Robles V, Grötsch H, et al. 897 Crosstalk between PI(4,5)P2 and CK2 Modulates Actin Polymerization during Endocytic 898 Uptake. Dev Cell. 2014 Sep;30(6). .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 36 899 900 Supporting Information 901 S1 Table. List of yeast strains used in this study. 902 S2 Table. List of plasmids used in this study. 903 S1 Fig. Rgf1 is required for proper localization of the Tea1-Tea4 complex at the cell tip. 904 (A) Wild-type cells expressing rgf1-GFP were stained with Calcofluor white (20 μg/mL) to spot 905 areas of growth. The arrows indicate the localization of Rgf1 at the growing tip in monopolar 906 cells. (B) Quantitation of the number of Tea4 dots associated to the MTs in WT and rgf1Δ strains 907 per cell. The mean ± SD of > 80 cells is shown. (C) The wild-type and rgf1Δ cells expressing tea4- 908 GFP were treated with the translation inhibitor cycloheximide (CHX, 100 μg/mL) for the 909 indicated times. Proteins were visualized by western blot with antibodies against GFP (Tea4) or 910 tubuline (Tub), as a loading control (upper). The graphic represents the quantification of Tea4 911 levels at different times (hours) after the treatment relative to time 0, which was assigned a 912 value of 1 (bottom). (D) The graphics show the MT polymerization (left) and depolymerization 913 (right) rates in wild-type and rgf1Δ cells. The mean ± SD of > 75 cells is shown. Statistical 914 significance was calculated using two-tailed unpaired Student’s t test. *P < 0.05; ****P < 0.0001; 915 ns= non-significant. 916 S2 Fig. Rgf1 cooperates with Mod5 in Tea4 anchoring to the cellular poles. 917 (A) Morphology of the wild-type, rgf1Δ, mod5Δ and rgf1Δmod5Δ strains after refeeding 918 treatment (-MBC). Scale bar 2 μm. (B) Cell morphology of the indicated strains after 4 hours at 919 36°C in YES liquid medium. Scale bar 2 μm. 920 S3 Fig. Rgf1 interacts with the cell end marker Tea4 and binds to phosphatidylinositol-4- 921 phosphate through its PH domain. 922 Coprecipitation of Rgf1 and Tea1. Cell extracts from cells producing Tea1-GFP, Rgf1-HA, and 923 Tea1-GFP and Rgf1-HA were precipitated with GFP-trap beads and blotted with anti-HA or anti- .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 37 924 GFP antibodies (co-immunoprecipitation and immunoprecipitation). Western blot was 925 performed on total extracts to visualize total Tea1-GFP and Rgf1-HA levels (whole cell extracts). 926 S4 Fig. Tea4 accumulation at the cell ends depends on Rgf1 anchoring to the PM and Rho1 927 activation. 928 Representative images of Rgf1-GFP and Rgf1ΔPTTR-GFP localization. Cells were stained with 929 calcofluor white (20 μg/mL) to spot areas of growth. Scale bar 2 μm. 930 S5 Fig. Rgf1 is part of actin-dependent machinery that signals growth poles in the absence of 931 Tea1–Tea4 complex. 932 (A) Representative images of LifeAct-GFP (actin) localization in cells untreated or treated with 933 KCl 0.6 M, sorbitol 1.2 M, or 37°C (heat) for 1 hour. The maximum-intensity projection of six Z- 934 slides (0.5 μm step-size) of fluorescence is shown. (B) Quantitation of the T-shaped cells in the 935 tea1Δ mutant treated with DMSO (Unt.), KCl 0.6 M, sorbitol 1.2 M, or 37°C (heat) for 1 hour, 936 then washed and allowed to grow without the drug for 3 hours. The graph represents the mean 937 ± SD of > 200 cells from two independent experiments. (C) Quantitation of T-shaped wild-type, 938 tea1Δ, tea1Δ rgf1Δ, and tea1Δ rgf1ΔPTTR cells grown to log phase in YES liquid medium at 28°C. 939 The graph represents the mean ± SD of > 200 cells from two independent experiments. (D) Cells 940 expressing rgf1-GFP and crn1-GFP (actin patches) or mCherry-atb2 (microtubules) cultured 941 separately; mixed; and then treated with DMSO, LatA 100 μM, or MBC 50 μM for 15 minutes. 942 (E) LifeAct-mCherry (actin) and Rgf1-GFP localization in wild-type cells treated with KCl 0.6 M for 943 1 hour and then washed and allowed to grow without stress for the indicated times. The 944 maximum-intensity projection of four Z-slides (0.6 μm step-size) of fluorescence is shown. 945 Statistical significance was calculated using a two-tailed unpaired Student’s t test. **P < 0.01; 946 ****P < 0.0001. Scale bar 2 μm. 947 S1 Movie. Tea4 and microtubule dynamics in wild-type cells. Tea4-GFP (green) and mCherry- 948 Atb2 (red) localization in wild-type cells. Protein dynamics was followed for 8 minutes, with .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint 38 949 pictures taken every 20 seconds. The maximum-intensity projection of six Z-slides (0.5 µm step- 950 size) is shown. 951 S2 Movie. Tea4 and microtubule dynamics in rgf1Δ cells. Tea4-GFP (green) and mCherry-Atb2 952 (red) localization in rgf1Δ cells. Protein dynamics was followed for 8 minutes, with pictures taken 953 every 20 seconds. The maximum-intensity projection of six Z-slides (0.5 µm step-size) is shown. 954 S3 Movie. Rgf1 and actin dynamics in wild-type cells during osmotic stress recovery. Rgf1-GFP 955 (green) and LifeAct-mCherry (actin in red) localization in wild-type cells treated with KCl 0.6 M 956 for 1 hour and then washed and allowed to grow without stress. Protein dynamics was followed 957 for 42 minutes, with pictures taken every 3 minutes. The maximum-intensity projection of four 958 Z-slides (0.6 µm step-size) is shown. 959 S4 Movie. Rgf1 and actin dynamics in tea1Δ cells during osmotic stress recovery. Rgf1-GFP 960 (green) and LifeAct-mCherry (actin in red) localization in tea1Δ cells treated with KCl 0.6 M for 1 961 hour and then washed and allowed to grow without stress. Protein dynamics was followed for 962 66 minutes, with pictures taken every 3 minutes. The maximum-intensity projection of four Z- 963 slides (0.6 µm step-size) is shown. 964 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 11, 2024. ; https://doi.org/10.1101/2024.01.10.574961doi: bioRxiv preprint