Portal de Periódicos da CAPES

Fission yeast IQGAP arranges actin filaments into the cytokinetic contractile ring

2009; Springer Nature; Volume: 28; Issue: 20 Linguagem: Inglês

10.1038/emboj.2009.252

1460-2075

Masak Takaine, Osamu Numata, Kentaro Nakano,

Plant Reproductive Biology

Article27 August 2009free access Fission yeast IQGAP arranges actin filaments into the cytokinetic contractile ring Masak Takaine Corresponding Author Masak Takaine Department of Structural Biosciences, Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan Search for more papers by this author Osamu Numata Osamu Numata Department of Structural Biosciences, Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan Search for more papers by this author Kentaro Nakano Corresponding Author Kentaro Nakano Department of Structural Biosciences, Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan Search for more papers by this author Masak Takaine Corresponding Author Masak Takaine Department of Structural Biosciences, Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan Search for more papers by this author Osamu Numata Osamu Numata Department of Structural Biosciences, Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan Search for more papers by this author Kentaro Nakano Corresponding Author Kentaro Nakano Department of Structural Biosciences, Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan Search for more papers by this author Author Information Masak Takaine 1, Osamu Numata1 and Kentaro Nakano 1 1Department of Structural Biosciences, Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan *Corresponding authors. Department of Structural Biosciences, Graduate School of Life and Environmental Sciences, University of Tsukuba, 1-1-1 Tennohdai, Tsukuba, Ibaraki 305-8577, Japan. Tel.:/Fax: +81 029 853 4530; E-mail: [email protected] or Tel.:/Fax: +81 029 853 6642; E-mail: [email protected] The EMBO Journal (2009)28:3117-3131https://doi.org/10.1038/emboj.2009.252 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The contractile ring (CR) consists of bundled actin filaments and myosin II; however, the actin-bundling factor remains elusive. We show that the fission yeast Schizosaccharomyces pombe IQGAP Rng2 is involved in the generation of CR F-actin and required for its arrangement into a ring. An N-terminal fragment of Rng2 is necessary for the function of Rng2 and is localized to CR F-actin. In vitro the fragment promotes actin polymerization and forms linear arrays of F-actin, which are resistant to the depolymerization induced by the actin-depolymerizing factor Adf1. Our findings indicate that Rng2 is involved in the generation of CR F-actin and simultaneously bundles the filaments and regulates its dynamics by counteracting the effects of Adf1, thus enabling the reconstruction of CR F-actin bundles, which provides an insight into the physical properties of the building blocks that comprise the CR. Introduction The bundling and crosslinking of actin filaments are crucial for organizing three-dimensional actin cytoskeletons, which are in contrast with the thin filaments in muscle cells in which the filaments align one-dimensionally and are discrete from each other. Cells use a variety of actin-binding proteins (ABPs) to organize single actin filaments into bundled, crosslinked, or branched structures to perform cellular processes, such as motility, adhesion, vesicle trafficking, and division. During the cytokinesis of many eukaryotes, a contractile ring (CR) is formed in the equatorial cortex and is believed to divide the cell in two by actomyosin-based contraction (Satterwhite and Pollard, 1992). The accumulation of F-actin and myosin II in the CR has been confirmed in many types of cells by fluorescence microscopy (FM). In some cells, the CR was also examined by electron microscopy (EM), which showed the bundling of F-actin (Sanger and Sanger, 1980; Yasuda et al, 1980; Mabuchi et al, 1988; Kamasaki et al, 2007). It should be underlined that the molecular entity responsible for the bundling of CR actin filaments remains to be determined. The actin-crosslinking protein α-actinin accumulates in the cleavage furrow of animal cells (Fujiwara et al, 1978; Mukhina et al, 2007) and is localized to the CR of fission yeast cells (Wu et al, 2001). Although mammalian and fission yeast α-actinins are involved in re-arrangement of the actin cytoskeleton during cytokinesis, their functions in CR formation seem subsidiary. Fimbrin and eukaryotic translation elongation factor 1A can bundle F-actin and are also present in the division furrow of Tetrahymena cells (Numata et al, 2000); however, their functions in CR formation are still unknown. Fimbrin also forms a medial ring during mitosis in fission yeast cells, but the ring does not contract and is not required for cytokinesis (Nakano et al, 2001; Wu et al, 2001). To date, the fission yeast Schizosaccharomyces pombe is the best model system for studying CR formation and cytokinesis. Molecular genetics has identified >50 genes involved in cytokinesis (Balasubramanian et al, 2004). Many accumulate in the CR and their temporal order (Wu et al, 2003), and local concentrations (Wu and Pollard, 2005) were determined by FM using green fluorescent protein (GFP) fusion proteins. Specifically, the incorporation of actin filaments into the CR was shown to occur in two phases: (1) accumulation of fine F-actin bundles in a mesh-like fashion and (2) actomyosin-driven compaction of the bundles into a tight ring (Arai and Mabuchi, 2002; Vavylonis et al, 2008). ABP, including tropomyosin Cdc8 (Balasubramanian et al, 1992), profilin Cdc3 (Balasubramanian et al, 1994), formin Cdc12 (Chang et al, 1997), IQGAP Rng2 (Eng et al, 1998), the actin-depolymerizing factor Adf1 (Nakano and Mabuchi, 2006), and myosin II Myo2 (Kitayama et al, 1997) localize to the CR and are required for CR formation in fission yeast. Taken together, the following picture of CR assembly emerges: in early mitosis, Cdc12 and Cdc3 nucleate and grow de novo actin filaments (Kovar et al, 2003), which are composed of a loose mesh, and then myosin II (Vavylonis et al, 2008) and Rng2 (Eng et al, 1998) organize this mesh into a ring. Adf1 is believed to (1) provide a source of G-actin, which is required for de novo actin ring formation, by breaking actin patches and (2) promote the turnover of actin in the CR (Nakano and Mabuchi, 2006). The functions of Cdc3 (Lu and Pollard, 2001; Takaine and Mabuchi, 2007), Cdc8 (Skoumpla et al, 2007; Skau et al, 2009), Cdc12 (Kovar et al, 2003), and Adf1 (Andrianantoandro and Pollard, 2006) in actin assembly were biochemically examined, whereas those of Rng2 and Myo2 remain to be elucidated. To understand the mechanisms of actin filament bundling in the CR of fission yeast, we attempted to elucidate the function of Rng2 in this study. By close observation of temperature-sensitive rng2 mutant cells, Rng2 was found to be involved in both the generation (i.e. nucleation and/or elongation) of CR F-actin and its organization into a ring. Domain analysis showed that a short N-terminal fragment of Rng2 is required for these functions and is localized to CR F-actin. Moreover, we biochemically examined the fragment and found that it promotes actin polymerization and bundles F-actin. Finally, we shed some light on the physical properties of the bundles, which seemed to be flexible, consisting of subdividable arrays of F-actin under EM, and were able to counteract Adf1-induced disassembly. These results suggest that Rng2 generates bundles of actin filaments that are used to produce the CR. Results Rng2 is involved in both the generation and convergence of F-actin, which comprises the CR Reportedly, temperature-sensitive rng2 mutant cells (rng2ts cells) show defects in the arrangement of F-actin into a ring after growth for 2–4 h at 36°C (restrictive temperature), displaying disorganized actin cables in the middle of the cell instead of a compact ring (Chang et al, 1996; Eng et al, 1998). We followed changes in the actin cytoskeleton of mutant cells for 2 h. In rng2ts cells at 25°C (permissive temperature), an F-actin ring was formed in the middle of cells (0 min in Figure 1A). Thirty minutes after the shift to 36°C, a trace of F-actin had accumulated in the middle of some binucleate cells instead of a ring, and grew to wrap randomly around the equator of the cell over 90 min (Figure 1A). Wild-type cells had normal medial F-actin rings at 25 and 36°C (Figure 1B). The increase in rng2ts cells without normal F-actin rings coincided with an increase in binucleate rng2ts cells (Figure 1C and D). Figure 1.Temperature-sensitive defects in actin ring formation of rng2 mutant cells. (A, B) After growing at 25°C, rng2ts (A) and wild-type (B) cells were shifted to 36°C and fixed after 0, 30, 60, or 90 min incubation before being stained for DNA and F-actin. Arrowheads indicate abnormal actin assembly in the medial region. (C) Percentage of binucleate cells. (D) Percentage of binucleate cells without normal CR. As these data include late telophase and septating cells, the CR was not found in about 20% of wild-type cells. Data are expressed as the means±s.d. (error bars) of three independent experiments. More than 120 cells were counted for each point. Note that rng2ts cells tended to aggregate, which prevented their correct observation, so we analysed cells outside of aggregates. (E, F) Phenotypes of rng2ts mutant cells at anaphase. Wild-type (E) or rng2ts (F) cells expressing GFP-atb2 (α-tubulin) were fixed and then stained for DNA and F-actin after 30 min incubation at 36°C. Cells with a long mitotic spindle were photographed. The percentages of cells with the depicted phenotypes are shown. About 40 cells from two independent experiments were observed for each strain. Asterisks and an arrowhead indicate the CR and the abnormal accumulation of F-actin, respectively. (G, H) Localization of Rlc1-GFP in the CR. After growing at 25°C, wild-type (G) or rng2ts (H) cells expressing Rlc1-GFP were fixed and then stained for DNA and F-actin. The percentages of cells with the depicted phenotypes are shown. More than 60 cells from two independent experiments were observed for each strain. Bars: 5 μm. Download figure Download PowerPoint We examined further when the phenotype appeared during mitosis. Most wild type and rng2ts cells with long mitotic spindles, indicating that they were in anaphase, had F-actin rings in their centre at 25°C (data not shown). Even at 36°C, most wild-type cells in anaphase still had equatorial F-actin rings (asterisks in Figure 1E). In contrast, 62% of rng2ts cells showed no accumulation of F-actin (Figure 1F), and 36% showed abnormal accumulation of F-actin (arrowhead in Figure 1F) at 36°C, indicating that the generation of CR F-actin was disturbed in anaphase. Even at 25°C, the fraction of binucleate rng2ts cells was significantly higher than that of wild-type cells (28±3% versus 19±1%, Figure 1C), suggesting delayed CR assembly or contraction in rng2ts cells because of the weakened function of Rng2. Thus, we examined the localization of Rlc1 (a light chain of myosin II) to the CR at 25°C. In wild-type cells at mitosis, Rlc1-GFP was observed in rings with uniform fluorescence intensity, and the rings were superimposable on F-actin rings in all cases (Figure 1G). In rng2ts cells, however, Rlc1-GFP was sometimes (45%) observed in rings with uneven fluorescence intensity (Figure 1H), even though F-actin formed uniform rings in such cells. The uneven distribution of Rlc1 in the CRs was observed also in living rng2ts cells (Supplementary Figure S1). We sometimes observed aggregate(s) of GFP fluorescence inside an Rlc1 ring in mutant cells. We also observed the actin cytoskeleton of wild type and rng2-null cells after spore germination (Supplementary Figure S2A and B). In cells derived from wild-type spores, a medial F-actin ring was formed at mitosis. In contrast, as described in the earlier study (Eng et al, 1998), a spot of F-actin was formed instead of the CR in rng2-null cells. In these cells, cytokinesis and septation were blocked, whereas polar growth and nuclear division progressed. The F-actin spot formed once at the first mitosis in the medial region and remained there. These results suggest that Rng2 is involved in the generation of medial F-actin in early mitosis, in the arrangement of the filaments into a ring before telophase, and in the uniform distribution of Rlc1 in CR. Rng2 interacts with F-actin through its N-terminal region Human IQGAP1 (Figure 2A, Hs IQGAP1) contains six protein-interacting domains: a single calponin homology domain (CHD, one of the common F-actin-binding domains (ABDs)) at its N-terminal; four coiled coil regions (CCs), which mediate dimerization (Fukata et al, 1997), a WW domain in its middle; a single CC domain, a Ras GTPase-activating protein (GAP)-related domain (GRD); and a RasGAP-C homology domain (RasGAP-C) in its C-terminal region. Although Hs IQGAP1 is a mammalian orthologue of Rng2, Rng2 lacks the four CCs and the WW domain: seven IQ motifs directly follow the CHD, and we failed to detect any CC in the N-terminal half of Rng2 (residues 1–745) using the COILS program (Lupas et al, 1991). Thus, the N-terminal region of Rng2 seems to have specific functions that are different from those of Hs IQGAP1. In this study, we sometimes compared Rng2 with fission yeast fimbrin, Fim1, as a control actin-bundling protein. Fim1 can bundle F-actin using two ABDs, each of which is composed of two CHD in tandem (Nakano et al, 2001; Figure 2A). Figure 2.Overexpression of proteins containing the N-terminal region of Rng2 produced aberrant actin cables and blocked CR formation. (A) The domain organizations of human IQGAP1, S. pombe Rng2, truncated proteins of Rng2, and S. pombe fimbrin Fim1. White numbers show relative amino-acid sequence homology in percent. (B, D, F, G) Phenotype of cells overexpressing Rng2 or its truncated proteins or (C) Fim1. (E) Mock overexpression was performed using an empty vector. Constructs were expressed from plasmids under the control of the strongest nmt1 promoter in wild-type cells (B–F) or cells expressing Sid4-mRFP (G). Expression was induced for 15 h at 30°C, and then the cells were processed for FM as described in Figure 1. Overlaid images of DNA (red) and F-actin are shown in (D–F). In (G), the RFP signals and DNA staining of living cells are shown. Bars: 5 μm. Download figure Download PowerPoint We observed the effects of the overexpression (OE) of Rng2 constructs (Figure 2A). OE of Rng2 (amino-acid residues 1–1489), Rng2N (residues 1–803), or Rng2Ns (residues 1–300) produced numerous thick actin cables, prevented CR formation, and dispersed actin patches over the cell cortex (Figure 2B; data for Rng2N is not shown). In these cells, cytokinesis and septation were blocked, whereas polar growth and mitosis were progressed. Although Rng2 seems to interact specifically with CR F-actin at mitosis (Eng et al, 1998), abnormally thick F-actin bundles were induced by OE of Rng2, Rng2N or Rng2Ns also in cdc25-22 cells arrested at the G2/M boundary (Supplementary Figure S3A). Thus, it was suggested that the bundling effects of these constructs are not necessarily specific to mitosis and to CR F-actin when they are overexpressed. In contrast, OE of Fim1 also prevented the formation of F-actin rings but produced abundant actin patches (Figure 2C). OE of Rng2ΔNs (residues 301–1489) did not disturb mitosis but specifically blocked CR formation: F-actin accumulated in the middle of cells during mitosis, but failed to form rings (Figure 2D). We noted that the accumulated F-actin was eventually cleared (i.e. turned over) and cells failed to complete septation (thick bars in Figure 2D). Interphase cells had normal actin patches and cables but not abnormal F-actin structures (b in Figure 2D; data not shown), which indicated that Rng2ΔNs essentially does not interfere with interphase F-actin structures. Mock OE through the pREP1 vector had no effect on the actin cytoskeleton (Figure 2E). On the other hand, OE of Rng2C (residues 804–1489) blocked mitosis but did not disturb polar growth (Figure 2F). In these cells, actin patches were normally located at the cell ends, and actin cables and rings were absent. All mononucleate cells had only one spindle pole body (SPB) on the nucleus, which confirmed that they were held in interphase (Figure 2G). OE of Rng2GRD (residues 804–1205) or Rng2RasGAP-C (residues 1168–1489) caused no growth or morphological defect (data not shown). These results suggest that the N- and C-terminal halves of Rng2 have different functions and that the former can produce thick F-actin bundles through the Rng2Ns moiety. Moreover, complementation assays using rng2ts and rng2-null cells showed that the Rng2Ns moiety is indispensable for the function of Rng2 during cytokinesis (see Supplementary data; Supplementary Figures S2 and S4). Rng2 contains multiple localization domains We examined the localization of various GFP-tagged Rng2 constructs in wild-type cells, as summarized in Figure 3J. The localization of various constructs is described also in the Supplementary data. GFP-Rng2 localized to the CR as reported earlier (Figure 3A; Eng et al, 1998; Wu et al, 2003). In addition to the rings, we detected GFP-Rng2 dots, as observed by Eng et al (1998), in some cells with prolonged induction, and they partially colocalized with SPB (Supplementary Figure S5A). GFP-Rng2Ns also localized to the medial rings that were able to contract (Figure 3B). GFP-Rng2ΔNs tended to aggregate in the cytoplasm and formed a dim medial ring during mitosis in some cells (Figure 3C). We noted that the localization of GFP-Rng2Ns was different from that of GFP-Rng2CHD (residues 1–189), which nonspecifically labelled all whole F-actin structures (Wachtler et al, 2003; Figure 3D), and that GFP-Rng2Ns did not mark actin patches or cables, even when it was abundantly expressed with prolonged induction (data not shown). Figure 3.Localization of Rng2 and its truncated proteins. All constructs were expressed in wild-type cells from plasmids under the control of the strongest nmt1 promoter. We observed cells that expressed the GFP fusions before the effects of OE became apparent. Images of (A) GFP-Rng2, (B) GFP-Rng2Ns, (C) GFP-Rng2ΔNs, and (D) GFP-Rng2CHD. (E–G) Localization of GFP-Rng2, GFP-Rng2ΔNs, and GFP- Rng2Ns in the absence of F-actin. Cells expressing each construct were treated with 0.2% DMSO alone (–Lat-A) or with 0.2% DMSO plus 4 μM Lat-A (+Lat-A) for 10 min at 25°C, before being stained for DNA. (H) Images of GFP-Rng2 or GFP-Rng2ΔNs expressed at a low level in the presence of 5 μM thiamine (repressed conditions). (I) Cells bearing pREP1-GFP-Rng2ts were grown under repressed conditions at 25°C, shifted to 25 or 36°C for 30 min, and stained for DNA. Arrowheads and small bars indicate the CR and bands of dots, respectively. Double arrowheads indicate the CR during contraction. (J) Summary of the localization of Rng2 constructs. Bars: 5 μm. Download figure Download PowerPoint We verified whether the assembly of these rings depends on F-actin using Latrunculin A (Lat-A), which completely disassembled all F-actin structures within 10 min whereas DMSO, the solvent for Lat-A, did not affect F-actin organization (data not shown). In Lat-A, GFP-Rng2 or GFP-Rng2ΔNs formed a band of dots or a discontinuous ring at the division site (+Lat-A in Figure 3E and F). In contrast, the rings of GFP-Rng2Ns (+Lat-A in Figure 3G) and the localization of GFP-Rng2CHD (data not shown) were completely cleared by Lat-A. DMSO did not affect the localization of the proteins tested (−Lat-A in Figure 3E–G). GFP-Rng2 and GFP-Rng2ΔNs formed similar equatorial bands of dots even in the absence of Lat-A when they were expressed at lower levels (Figure 3H), whereas other constructs did not form such bands under any conditions (data not shown). GFP-Rng2C accumulated around the nucleus and formed a single dot in both interphase and mitotic cells (Supplementary Figure S5C and D) independently of F-actin (data not shown). The dots were multiplied in some cells and colocalized with SPB (Supplementary Figure S5E). GFP-Rng2C was also weakly localized to the medial region as a discontinuous band in septum-forming cells. These observations show that Rng2 (1) accumulates at the division site through Rng2ΔNs independently of F-actin and (2) probably associates with CR F-actin through Rng2Ns, and (3) the C-terminal half of Rng2 alone is able to accumulate in the perinucleus, including around SPB and the septum. We also addressed the localization of temperature-sensitive mutant Rng2 protein in wild-type cells. We analysed the DNA sequence of the rng2-D5 allele and identified a single mutation that substitutes the relatively conserved glycine residue at position 1032 for glutamic acid (Supplementary Figure S4A). The GFP-tagged mutant Rng2 (GFP-Rng2ts) formed dim medial rings during cytokinesis at 25°C (Figure 3I). We failed to find GFP-Rng2ts forming an equatorial band of dots, and the rings of GFP-Rng2ts disappeared in Lat-A (Figure 3J; data not shown), indicating that mutant Rng2 forms an F-actin-dependent medial ring. At 36°C, the rings of GFP-Rng2ts completely disappeared within 30 min (Figure 3I), whereas those of GFP-Rng2 remained intact (data not shown). Thus, the mutant Rng2 is unable to interact with CR F-actin at 36°C for some unknown reason, which causes defects in CR formation in rng2ts cells (Figure 1). Rng2Ns bundles F-actin To elucidate how Rng2 interacts with F-actin during CR formation, we considered two constructs, Rng2Ns and Rng2CHD. N-terminally His6-tagged Rng2Ns or Rng2CHD was expressed in bacteria and successfully purified (Figure 4A). In gel filtration chromatography, Rng2Ns was eluted as a single peak (Figure 4B), and the apparent molecular weight was estimated to be 41k (Figure 4C), which was only 10% larger than the deduced molecular weight of the Rng2Ns monomer (37k). Thus, Rng2Ns appeared monomeric at an ionic strength of 0.05 M and a pH of 7.5. Figure 4.Cooperative association of Rng2Ns or Rng2CHD with F-actin. (A) Purified proteins resolved by SDS–PAGE and stained with CBB. (B) Gel filtration chromatography of Rng2Ns. Arrowheads indicate the peak elution volumes of the following molecular weight markers from left to right: Blue dextran 2000 (void volume), BSA (68k), ovalbumin (45k), chymotrypsinogen A (25k), and RNase A (14k). (C) The peak elution volume of Rng2Ns was plotted on a calibration line as a square, giving an apparent molecular weight of 41k. (D, E) High- (D) or low-speed (E) co-sedimentation of Rng2Ns, Rng2CHD, or Fim1 (1.5 μM) with F-actin (3 μM). Equal portions of the supernatants (S) and pellets (P) were subjected to SDS–PAGE and CBB-stained. (F) The high-speed co-sedimentation assay shown in (D) was performed for various concentrations of Rng2Ns, Rng2CHD, or Fim1 (0.3–3.5 μM). The concentration of the bound Rng2Ns (•), Rng2CHD (⧫), or Fim1 (▴) was plotted against the concentration of the free portion and fitted with a hyperbolic function, and representative plots are shown. (G) The low-speed co-sedimentation assay shown in (F) was performed for various concentrations of Rng2Ns, Rng2CHD, or Fim1. The amount of actin pelleted was plotted against the Rng2Ns/actin, Rng2CHD/actin, or Fim1/actin concentration ratio. Data are expressed as the means±s.d. (error bars) of three independent experiments. (H) Summary of the properties of the actin-bundling proteins tested. The dissociation constants (Kd) and stoichiometries were calculated from data obtained by three independent experiments. Download figure Download PowerPoint By high-speed sedimentation, in which F-actin mostly sedimented (Figure 4D, lane A), a significant amount of Rng2Ns, Rng2CHD, or Fim1 co-sedimented with F-actin (Figure 4D, lanes A+N, A+C, and A+F), whereas they barely sedimented at all in the absence of F-actin (Figure 4D, lanes N, C, and F). Quantitative analysis showed that Rng2Ns, Rng2CHD, and Fim1 bind with F-actin with Kd values of 0.88, 0.21, and 0.42 μM, respectively (Figure 4F). At saturation, the amount of bound Rng2Ns, Rng2CHD, and Fim1 per 1 mol actin was estimated to be 1.4, 1.4, and 1.6 mol, respectively. Next, we assessed whether Rng2Ns or Rng2CHD can crosslink F-actin using a low-speed sedimentation assay in which F-actin sediments only when it aggregates (Figure 4E, lane A). In the presence of Rng2Ns, Rng2CHD, or Fim1, a considerable amount of F-actin was recovered in the pellet (Figure 4E, lanes A+N, A+C, and A+F). The amount of F-actin sedimented showed a sigmoidal dependence on the Rng2Ns or Rng2CHD concentration, suggesting that they crosslinked the filaments in a cooperative manner (Figure 4G). Rng2Ns sedimented F-actin more effectively than Rng2CHD. Meanwhile, F-actin was sedimented with hyperbolic dependence on the Fim1 concentration, indicating its non-cooperative crosslinking of F-actin. The molecular ratios of Rng2Ns, Rng2CHD, or Fim1 to actin that gave half-maximal sedimentation were 0.2, 0.4, and 0.03, respectively. Figure 4H summarizes the properties of the actin-bundling proteins. We examined the aggregates of F-actin formed by Rng2Ns, Rng2CHD, or Fim1 by FM. In their presence, thick bundles of rhodamine–phalloidin-labelled F-actin were detected (insets in Figure 5A–C), whereas only single filaments were observed for F-actin alone (Figure 5D). Bundles formed by Rng2Ns (Rng2Ns/actin bundles) were evenly distributed in the field, and some formed radial arrays (asterisks in Figure 5A). The bundles formed by Fim1 (Fim1/actin bundles) were often entangled and formed larger clots (Figure 5B), which may indicate that Fim1 can crosslink the filaments at various angles. Some single filaments were still present in the presence of Rng2Ns or Fim1 (dim filaments in the background of Figure 5A and B). The bundles formed by Rng2CHD were often loose (arrowheads in Figure 5C), and numerous single filaments always coexisted around them, suggesting that bundling by Rng2CHD was weaker than by Rng2Ns or Fim1. We further examined the ultrastructures of the bundles by EM. Rng2Ns/actin bundles appeared to be both thick bundles and radial arrays under EM (Figure 5E and F). Fim1/actin bundles seemed rather disentangled under EM: single bundles stood out, and large clots were rarely seen (Figure 5G). Only single filaments with a mean width of 6.9±1.2 nm were observed for F-actin alone (Figure 5H). Rng2Ns/actin bundles were more closely packed than Fim1/actin bundles: individual filaments in Fim1/actin bundles were distinguishable from each other, whereas those in Rng2Ns/actin bundles were obscure (Figure 5E′–H′). The mean width of Rng2Ns/actin bundles was 26±8 nm, whereas that of Fim1/actin bundles was distributed more widely with a mean of 36±17 nm (Figure 5I). Unlike Fim1/actin bundles, Rng2Ns/actin bundles were curvy, often smoothly split (Figure 5J), and sometimes twisted (Figure 5K), suggesting that they were flexible and could be divided into smaller units. Figure 5.Rng2Ns bundles F-actin. (A–D) Fluorescence micrographs of actin bundles. G-actin (2 μM) was polymerized alone (D), or with 0.6 μM Rng2Ns (A), Fim1 (B), or 1.8 μM Rng2CHD (C) for 10 min (A) or 50 min (B, C), and then stained with rhodamine–phalloidin. Asterisks indicate the centres of the radial arrays of actin bundles. Arrowheads indicate points where the bundles were loosened. (E–H) Electron micrographs of actin bundles. (I) Width distribution of Rng2Ns/actin (black bars) and Fim1/actin (white bars) bundles. The means±s.d. are depicted in the graph. The widths of the bundles were measured at more than 170 points for each condition. (J) Rng2Ns/actin bundles splitting. (K) Rng2Ns/actin bundles twisting (a) and simultaneously splitting (b). Bars: 10 μm (A–D) or 200 nm (E–H, J, and K). Download figure Download PowerPoint Rng2Ns accelerates spontaneous actin polymerization From the result suggesting the involvement of Rng2 in the generation of CR F-actin (Figure 1), we inferred that Rng2 might have an effect on the kinetics of actin polymerization. Rng2Ns promotes actin polymerization by decreasing the initial lag because of nucleation (Figure 6A). The effect of Rng2Ns became apparent at a more than 1:100 molar ratio to actin, and was comparable to that of the activated fission yeast Arp2/3 complex as a positive control (Figure 6F) at higher concentrations. In contrast, Fim1 decelerated actin polymerization at a more than 1:10 molar ratio to actin. It is noteworthy that Rng2Ns/actin bundles emerged in 10 min (Figure 5A) when Rng2Ns-promoted polymerization was nearly complete, indicating that in the presence of Rng2Ns, actin polymerization and F-actin bundling occur simultaneously. Figure 6.Rng2Ns accelerates spontaneous actin polymerization. (A) Time course of the polymerization of 2 μM Mg-G-actin (1% pyrene labelled) in the presence of Rng2Ns, Fim1 or fission yeast Arp2/3 complex (10 nM, activated by 67 nM GST-N-WASp-VCA). (B) Time course of the polymerization of 2 μM Mg-G-actin in the presence of Rng2Ns and Cdc3. Dashed and solid lines indicate samples without and with Cdc3 (5.3 μM), respectively. The corresponding time courses until 3000 s are shown in Supplementary Figure S6. (C) The time course of the polymerization of 2 μM Mg-G-actin in the presence of Rng2CHD. Inset shows the curves normalized to fluorescence intensities at the plateau. (D) The time to half-maximal polymerization