The kinesin-like protein TOP promotes Aurora localisation and induces mitochondrial, chloroplast and nuclear division

Almost all eukaryotic cells in the plant kingdom possess three kinds of organelles that contain DNA and have double membranes: one nucleus, many mitochondria and many chloroplasts (plastids). Since mitochondria and chloroplasts were derived from free-living α-proteobacterial and cyanobacterial ancestors, respectively, they are never synthesized de novo and their continuities are maintained by division, as is the nucleus. The cell cycle, therefore, consists of not only the mitotic cycle but also mitochondrial and chloroplast division cycles (Suzuki et al., 1994; Imoto et al., 2010). In plant cells, mitochondrial and plastid DNA replications take place before nuclear DNA replication (Kobayashi et al., 2009; Imoto et al., 2010). Recently, Kobayashi et al. have shown in Cyanidioschyzon merolae and tobacco BY-2 cells that mitochondrial and plastid DNA replications are signaled by the intracellular accumulation of a tetrapyrrole intermediate, probably Mg-ProtoIX, resulting in the activation of cyclin-dependent kinase A (CDKA, also known as Cdc2 in fission yeast) and the consequent initiation of nuclear DNA replication (Kobayashi et al., 2009). Therefore, it seems that an integrated control system that induces division of the three types of double-membrane organelles may be hidden in the initiation step of mitochondrial and plastid divisions.

In the last two decades, it has been shown that mitochondrial and chloroplast divisions occur in three steps: formation of mitochondrial-dividing (MD) machinery and plastid-dividing (PD) machineries at each division site (Kuroiwa et al., 1998), constriction of the division site, and pinching-off of the bridge of the daughter organelle (Fig. 1A). During the formation step, MD and PD machineries are connected with each other and form a complex structure (MD/PD machinery complex), but this complex separates in the constriction step (Fig. 1A) (Yoshida et al., 2009). Both MD and PD division machineries comprise a chimera of inner rings of bacterial-derived proteins, such as FtsZ (Osteryoung and Nunnari, 2003; Kuroiwa et al., 2008) and eukaryote-specific proteins, such as the MD ring, PD ring and dynamin proteins (Bleazard et al., 1999; Miyagishima et al., 2003; Gao et al., 2003; Nishida et al., 2003; Osteryoung and Nunnari, 2003; Kuroiwa et al., 2008). Recent studies showed that dynamin is localized between the PD ring filaments and is essential for the generation of the motive force for contraction (Yoshida et al., 2006). In addition, the PD ring is constructed of a bundle of glycosyltransferase protein PDR1-mediated-polyglucan filaments. Thus, the contraction of the PD machineries is caused by the sliding movement between dynamin and polyglucan filaments (Yoshida et al., 2010). Similarly, it has been thought that the contraction of the MD machineries is probably driven in the same way as that of the PD machinery (Nishida et al., 2003; Yoshida et al., 2009; Kuroiwa et al., 2008).

In addition, it has been reported that in C. merolae a pre-spindle structure, designated as the mitochondrial spindle, is formed from each spindle pole to the division site of mitochondria before nuclear division (supplementary material Fig. S1, white arrowheads) (Nishida et al., 2005; Imoto et al., 2010). Similar to the cell division spindle (Hirokawa, 1998; Walczak and Heald, 2008), the mitochondrial spindle may be also organized by tubulin-mediated microtubule polymers, several types of kinesin-superfamily proteins, and many relating factors such as mitotic kinases.

Our recent study showed that a mitotic serine/threonine kinase, Aurora kinase (AUR), which is encoded by a single-copy gene in the genome C. merolae, is localized not only at spindle poles and the cell division spindle but also the mitochondrial spindle (Kato et al., 2011). Specifically, AUR accumulates from the mitochondrial spindle to the mitochondrial division site in the contraction phase of mitochondrial division, indicating that it is involved in the activation of the MD machinery. Thus, these findings suggest the existence of an uncharacterized pathway or factor that coordinates timing of mitochondrial and chloroplast division and links mitochondrial- and chloroplast-division cycles with the cell-division cycle.

C. merolae offers advantages for studying the regulation of integrated initiation of nuclear, mitochondrial and chloroplast divisions. The cell contains just one chloroplast, one mitochondrion and one nucleus (Matsuzaki et al., 2004), the division of which occurs in that order in highly synchronized cells controlled by light/dark cycles (supplementary material Fig. S1) (Suzuki et al., 1994). In addition, availability of the complete genome sequence of C. merolae facilitates highly sensitive transcriptomic and proteomic analyses (Matsuzaki et al., 2004; Nozaki et al., 2007; Yagisawa et al., 2009; Fujiwara et al., 2009; Yoshida et al., 2009; Yoshida et al., 2010).

In this study, we revealed that a novel kinesin-like protein, TOP, induces divisions of the nucleus, mitochondrion and chloroplast. TOP localized on the MD/PD machinery complex to transfer AUR just before the contraction of the MD and PD machineries. By this process, the mitochondrion and chloroplast were divided by the MD and PD machineries, respectively. Nuclear division was also accomplished by the TOP-mediated spindle. Thus, TOP regulates division of three organelles, the nucleus, mitochondrion and chloroplast, by relocalization of AUR protein.

To identify proteins that regulate the integrated initiation of nuclear, mitochondrial and chloroplast division, we performed a proteomic analysis of the isolated MD/PD machinery complex from cells in early S phase (Fig. 1B; supplementary material Fig. S2; Table S1; see Materials and Methods) because division of the organelles and organization of the mitochondrial spindle begins immediately after this phase. For isolation of non-constricting MD/PD machinery complexes, synchronized cells were harvested at the start of the second dark period (S phase) and treated with Nonidet P-40 and n-octyl-β-D-glucopyranoside (see Materials and Methods). After isolation of the non-constricting MD/PD machinery complexes, proteomic analysis was performed using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS). In this fraction, known components of the MD and PD machinery, Dynamin 1 (Dnm1), Mda1 and Dnm2 were identified. Comparing previous results of proteomic analyses of the MD/PD machineries isolated from cells in M phase (Yoshida et al., 2009; Yoshida et al., 2010), we identified a novel S-phase-specific protein TOP (three-organelle divisions inducing protein; Fig. 1B). Using the completely sequenced genomic information of C. merolae (http://merolae.biol.s.u-tokyo.ac.jp/), we distinguished that TOP was one of the C-terminal motor domain-type kinesin-superfamily proteins [C-terminal KIFs (reviewed by Hirokawa et al., 2009)] (Fig. 2A), which had not previously been detected in isolated constricting MD and PD machineries from cells in M phase (Yoshida et al., 2009; Yoshida et al., 2010). By phylogenetic analysis, we showed that a kinesin motor domain of TOP is a classical C-terminal KIF, and TOP homologs are widely conserved throughout eukaryotic cells, especially plants (supplementary material Fig. S3). Some groups of C-terminal KIFs have been reported to transport organelles (Saito et al., 1997; Xu et al., 2002; Bananis et al., 2004; Hirokawa et al., 2009). Also, one of the known functions of C-terminal KIFs is the assembly of spindle poles by the cross-linking of parallel microtubules in each half-spindle, where they focus minus ends into spindle poles (Walczak et al., 1997; Hirokawa, 1998); however, the involvement of TOP in mitochondrial and/or chloroplast division has not previously been reported. We, therefore, examined the function of TOP to reveal the regulatory mechanism of nuclear, mitochondrial and chloroplast division.

The genes of known MD- and PD-machinery-associated proteins (Mda1 for MD machinery and Dnm2 for PD machinery) are selectively transcribed before organelle division; therefore, we examined the level of TOP transcription during the cell cycle using C. merolae microarray data (Fujiwara et al., 2009). The level of transcription of TOP increased during S phase and into early M phase, and this expression profile coincided with that of other known MD- and PD-machinery-associated proteins (Fig. 2B). We then generated anti-TOP antibodies via bacterially expressed TOP protein. Anti-TOP antibodies detected protein of the predicted molecular mass of TOP (∼96 kDa; Fig. 2C). Also, immunoblot analyses of total protein from synchronized cultures detected TOP during chloroplast, mitochondrial and nuclear divisions (Fig. 2D). Next, we examined the intracellular localization of TOP during the cell cycle by immunofluorescence microscopy using the anti-TOP antibody (Fig. 3). Although TOP was not detected during the G1 phase, it was observed in S phase, indicating that TOP was located near the mitochondrial division site (Fig. 3, white arrowhead). Then, in the S–G2 phases, the fluorescence signal of the TOP-assembled small spindle pole increased in size and split into two spindle poles (Fig. 3, black arrowheads). During M phase, TOP was localized on microtubules in addition to at the two spindle poles. Therefore, we conclude that TOP is involved in the spindle poles and the spindle microtubules for nuclear division in M phase.

Because TOP was identified in the fraction of the isolated MD/PD machinery complex by proteomic analysis (Fig. 1B) and appeared near the mitochondrial division site (Fig. 3, white arrowhead), we next investigated whether TOP was involved in mitochondrial and/or chloroplast division in addition to nuclear division. For the purpose, we examined the effects of TOP downregulation using antisense suppression (Fig. 4). Twenty-four hours after transformation, antisense TOP induced defects in mitochondrial and chloroplast divisions (P<0.001; Fisher's exact test; Fig. 4A,B). In these cells, both MD and PD machineries were found at the mitochondrial and chloroplast division sites, respectively, although constriction had not commenced in either (Fig. 4C,D). In contrast, inhibition of microtubule organization by oryzalin treatment arrested nuclear division of the cells, but both mitochondrial and chloroplast divisions were performed normally (supplementary material Fig. S4) (Nishida et al., 2005). In addition, immunofluorescence microscopy showed that a centromere marker protein, CENH3, appeared as multiple discrete speckles in nuclei of TOP-downregulated cells arrested in S/G2 phases (supplementary material Fig. S5). Therefore, the obstruction of mitochondrial and chloroplast divisions by TOP downregulation are not caused by spindle checkpoint. However, we cannot completely exclude the possibility that the S/G2 arrest was caused by activation of other checkpoints.

As TOP was required for the initiation of three-organelle divisions and seemed to be located near the mitochondrial division site, we examined whether the timing and manner of localization of TOP was related to the MD/PD machinery complex, using immunofluorescence microscopy. Both MD and PD machineries were formed prior to the expression of TOP (Fig. 5A). TOP seemed to be located only on the MD machinery in vivo (Fig. 5A) but could be recognized on the division site of the isolated dividing mitochondria and chloroplasts (Fig. 5B). To reveal whether TOP binds directly to the MD machinery complex, we isolated the MD/PD machinery complexes from cells in S or M phases and examined the localization of TOP by immunofluorescence microscopy (Fig. 6A). We also measured the circumference of isolated PD machineries to detect the stage of division, because the circumference of the PD machinery reduces as chloroplast division progresses (Miyagishima et al., 2001). The average circumference of the PD machineries containing TOP was 2.6 µm (n = 25; supplementary material Fig. S6), therefore, we concluded that TOP bound to the MD/PD machinery complexes that were derived from cells in the early phase of mitochondrial and chloroplast divisions. However, TOP did not associate with small MD/PD machinery complexes derived from cells in late stages of mitochondrial and chloroplast division (Fig. 6A). These results were also confirmed by immuno-electron microscopy (immuno-EM; Fig. 6B,C). Immunogold particles indicating TOP were distributed in a spot-like pattern on the MD machinery (Fig. 6C, inset). These results indicate that the direct interaction between TOP and larger MD machineries plays an important role in the initiation of constriction of MD and PD machineries.

Reports thus far indicate that protein phosphorylation is involved in at least the mitochondrial division. The WD40 protein Mda1 together with the MD ring assembles the outer ring of the MD machinery, and then Mda1 is phosphorylated. In the contraction phase of the mitochondrial division, phosphorylated Mda1 oligomers on the MD machinery are disassembled by GTP hydrolysis of Dnm1 (Nishida et al., 2007). And finally, the mitochondrial dynamin ring, mediated by Dnm1, severs the bridge of the dividing mitochondrion (Nishida et al., 2003). Since phosphorylation of Mda1 is induced with the progress of the mitochondrial division phases, protein phosphorylation is likely to mediate changes in the organization of the MD machinery for constriction. Thus, it is thought that phosphorylation is one of the most important modifications for the MD and PD machineries. Based on these results and hypothesis, we examined whether proteins in the isolated MD/PD machinery complexes were phosphorylated in the early and late phases of mitochondrial and chloroplast division, using cytochemistry and MALDI-TOF-MS. Large MD/PD machinery complexes, which were derived from the cells in early phase of division, had few fluorescence signals derived from phosphorylation, but small MD and PD machineries, which were derived from the cells in late phase of division, had strong fluorescence signals (Fig. 7A,B). Moreover, differences in fluorescence signals between the control fraction and the isolated MD/PD machinery complex fraction after staining for phosphorylated proteins showed that many proteins, including Dnm1, Mda1 (involved in mitochondrial division) and Dnm2 (involved in chloroplast division) in the isolated MD/PD machinery complex fraction, were phosphorylated (Fig. 7C,D). Phosphorylated peptides were estimated in matched peptide fragments of Dnm1, Mda1 and Dnm2 using MALDI-TOF-MS analyses (supplementary material Tables S2–S4). Both MD and PD machineries were highly phosphorylated in line with the progress of mitochondrial and chloroplast division. In addition, we showed that cells expressing antisense TOP contained the non-constricted MD and PD machineries at each division site (Fig. 4C), and that they were not phosphorylated (Fig. 7E). Lastly, we used immunofluorescence microscopy to confirm the kinases associated with the MD/PD machinery complex. Similar to C-terminal KIFs, it is known that Aurora kinase is also involved in spindle pole maturation because Aurora kinases were originally identified as being required for accurate spindle pole structure (Glover, et al., 1995; Andrews et al., 2003). In addition, a recent study showed that Aurora kinase in C. merolae (named AUR) was involved with not only mitotic spindle formation but also mitochondrial division (Kato et al., 2011). The level of transcription of AUR increased in S and M phases and this expression profile coincided with that of TOP (supplementary material Fig. S7). Therefore, we investigated whether AUR is involved with TOP during mitochondrial and chloroplast divisions. Immunofluorescence microscopy identified that AUR colocalized with TOP and was localized around the mitochondrion before the constriction phase of mitochondrial and chloroplast divisions (Fig. 7F, white arrowhead). Subsequently, AUR accumulated at the mitochondrial division site (Fig. 7F, black arrowhead). These results implied that AUR mediates phosphorylation of the MD and PD machineries and spindle pole maturation. Next, we investigated the localization of AUR in the TOP-downregulated cells to examine whether AUR is involved in the maturation of the MD and PD machineries. AUR protein was not found to be localized around the mitochondrial division site but scattered in the cytosolic region in the TOP-downregulated cells (Fig. 7G). In addition, the proteomic analysis of isolated MD/PD machinery complexes suggested that some of the phosphorylated peptides of Dnm1, Dnm2 and Mda1 were processed by Aurora kinase (supplementary material Tables S2–S4). In conclusion, TOP is a regulator of MD/PD machinery complexes for mitochondrial and chloroplast divisions, probably by transferring of Aurora kinase to the machinery.

Thus far, it has been revealed that mitochondrial and chloroplast division is performed by the MD and PD machineries which are large protein complexes. Although the molecular mechanisms involved are still not fully known, recent studies showed that FtsZ and dynamin proteins can generate a force for contraction of membranes by GTP hydrolysis (Osawa et al., 2008; Mears et al., 2011). In particular, a series of analyses of isolated single PD and single MD machineries showed that Dnm2 and Dnm1 are required to generate constriction force by sliding movements of the PD or MD ring filaments, respectively (Yoshida et al., 2006; Yoshida et al., 2009). In addition, it was also revealed that phosphorylation of Mda1 may be required to change the conformation of Dnm1 in the MD machinery to constrict the mitochondrial division site (Nishida et al., 2007). Thus, protein modification of the components of the MD and PD machineries are essential to accomplish mitochondrial and chloroplast division. However, such a regulation factor for protein modification of the MD and PD machineries has not been identified so far. In this study, using proteomic analyses, we identified a kinesin-like protein TOP of isolated MD/PD machinery complexes before constriction (Fig. 1B; supplementary material Fig. S2). Since immunofluorescence microscopy showed that during mitochondrial and chloroplast divisions TOP was localized on the MD/PD machinery complex just before constriction (Figs 5, 6), it was hypothesized that TOP is required for activation of the MD and PD machineries. Finally, a series of analyses of the antisense suppression of TOP (Fig. 4) and staining of phosphoproteins in the MD and PD machineries (Fig. 7) showed that TOP is required to induce contraction of the MD and PD machineries by phosphorylation. TOP is directly bound to the MD machinery in late S phase, then, TOP may induce activation of the MD/PD machinery complex through protein phosphorylation by transferring the mitotic kinase AUR from the cytosol to the MD/PD machinery complex (supplementary material Fig. S8). Combined with the results of downregulation of TOP (Fig. 4) and localization of TOP on the MD/PD machinery complex (Fig. 5), it is thought that mitotic kinases need to interact with a small part of the MD/PD machinery complex which is slightly exposed in the cytosol to activate the MD/PD machinery complex. Indeed, immunofluorescence signals of AUR increased in this part of the MD/PD machinery complex in line with the progresses of mitochondrial and chloroplast division (Fig. 7F). In order for the mitotic kinases to interact with the MD/PD machinery complex TOP may be required (supplementary material Fig. S8). Mitotic kinases associated with TOP can easily move to the periphery of the MD machinery and can also move to the periphery of the PD machinery along the MD machinery (supplementary material Fig. S8). Although we showed that Aurora kinase may be responsible for the MD and PD machineries, other kinases would be needed for the regulation of the machineries. The mammalian kinase CDK1, an orthologue of the plant CDKA, regulates phosphorylation of mitochondrial dynamin, Drp1, to enhance mitochondrial division during mitosis (Taguchi et al., 2007). CDKA and CDKB are cell cycle regulators in C. merolae (Kobayashi et al., 2011) and CDKB mRNA accumulation is specifically detected in the mitochondrial and chloroplast division phase of C. merolae (supplementary material Fig. S9). Thus, CDKA and CDKB in C. merolae are candidates to phosphorylate MD/PD machinery complexes in addition to Aurora kinase.

Previously, it was revealed that kinesin-superfamily members were important molecular motors that directionally transport various cargos, including single/double-bounded membranous organelles and large protein complexes (Hirokawa et al., 2009). Also, it was well known that kinesin proteins that are expressed in M phase are involved in the organization and function of the mitotic spindle (Walczak and Heald, 2008). These mitotic kinesins typically act in chromosome alignment and segregation. However, the involvement of a kinesin-like protein in mitochondrial and/or chloroplast division has never been reported, until now. Thus, this report is the first evidence that a member of the kinesin-superfamily proteins is involved in mitochondrial and chloroplast division. TOP induces protein phosphorylation of MD and PD machineries to accomplish mitochondrial and chloroplast divisions prior to nuclear division. And finally, nuclear division is performed by the TOP-mediated spindle structure (Fig. 3; supplementary material Fig. S8). Thereby, mitochondria and chloroplast division cycles are combined with the cell division cycle.

Recently, it was revealed that the nuclear division that is performed by TOP-mediated spindle poles (Fig. 3) is accompanied by the inheritance of the endoplasmic reticulum (ER) and Golgi apparatus in association with the mitotic spindle (summarized in supplementary material Fig. S10) (Imoto et al., 2011; Yagisawa et al., 2012a; Yagisawa et al., 2012b). Although the interaction of plant C-terminal KIFs with organelles is still unclear (Cai and Cresti, 2012), it was found that animal C-terminal KIFs are involved in intracellular transportation of these single-bounded membranous organelles. KIFC3 transports the Golgi apparatus and KIFC2 transports early endosomes as cargo (Saito et al., 1997; Hirokawa, 1998; Xu et al., 2002; Bananis et al., 2004; Hirokawa et al., 2009). Moreover, we showed that the initiation of both mitochondrial and chloroplast divisions is regulated by TOP, and recent studies revealed that the divided mitochondria are required as the carrier of microbodies and lysosomes, which are connected with divided mitochondria (Fujiwara et al., 2010; Imoto et al., 2010; Imoto et al., 2011). Thus, TOP is involved in the proliferation of all single- and double-membrane-bound organelles (supplementary material Fig. S10). In addition, the involvement of TOP in mitochondrial and chloroplast division, given the presence of TOP homologs in many members of eukaryotes (supplementary material Fig. S3), may indicate the original function of C-terminal kinesin-like proteins, which control proliferation of all double- and single-membrane-bound organelles.

The 10D strain of Cyanidioschyzon merolae was used (Matsuzaki et al., 2004), and was cultured in flasks, with shaking, under continuous light (40 W/m²) at 42°C. For synchronization, the cell cultures were subcultured to <1×10⁷ cells/ml in a flat-bottomed flask and subjected to a 12-hour light/12-hour dark cycle at 42°C using an automatic light/dark cycle CM incubator (Fujimoto Rika, Tokyo, Japan) (K. Suzuki et al., 1994). Synchronized cells were harvested at the start of the second dark period (S phase) or after 2 hours of the second dark period (M phase). Dividing chloroplasts and mitochondria and intact PD and MD machineries were isolated as described previously (Yoshida et al., 2009).

Samples were analyzed by a peptide mass fingerprinting (PMF) search using a MALDI TOF AXIMA TOF² mass spectrometer (Shimadzu, Kyoto, Japan) in reflectron mode. Database searches were performed using the software program MASCOT v2.2.01 (Matrix Science, MA, USA) running on the local server against the C. merolae genome database (including 5014 sequences) based on the FASTA file distributed by the Cyanidioschyzon merolae Genome Project (http://merolae.biol.s.u-tokyo.ac.jp/). The permissible value of missed cleavages was set at one. MS tolerance values were set at 0.2–0.4 Da. Identified proteins (a MASCOT score of more than 50) are listed in supplementary material Fig. S2; Table S1.

Additional amino acid sequences of C-terminal kinesin-superfamily members were obtained from GenBank (http://www.ncbi.nlm.nih.gov/genbank/) and were automatically aligned using CLUSTAL X, version 2.0.9 (http://www.clustal.org/download/current/) (Larkin et al., 2007). For phylogenetic analyses, ambiguously aligned regions were manually arranged or deleted using BioEdit Sequence Alignment Editor, version 4.8.10 (http://www.mbio.ncsu.edu/BioEdit/bioedit.html), resulting in 363 amino acids (including inserted gaps) being used. Phylogenetic tree construction and bootstrap analyses were performed using PHYLIP, version 3.66 (http://evolution.genetics.washington.edu/phylip.html), PROTDIST and NEIGHBOR for neighbor-joining (NJ), PROTPARS for maximum parsimony (MP), and PROML for maximum likelihood (ML). The JTT + Γ model (among-site rate variation model with four rate categories) was selected as the probability model for NL and ML analyses. Multiple datasets for bootstrap analyses (1000 replicates for NJ and MP, 100 replicates for ML) were calculated using CONSENSE. No outgroup was used to root the tree.

To generate an anti-TOP antiserum in guinea pig, amino acids 1–431 of the predicted 431-amino-acid sequence of the CMR497C protein was amplified by PCR using the following primers: 5′-cgggatccatgattcgagacagggttccag-3′ and 5′-cccaagctttgcgcgaagttccatgatc-3′. The resultant DNA fragment was cloned into pQE80L (Qiagen, Hilden, Germany) following restriction digestion at BamHI and HindIII sites. Protein expression and purification was performed as previously described (Nishida et al., 2005). To produce a protein expression vector for AUR, the full length of the coding sequence for AUR was amplified by PCR with the following primer: 5′-atggtaccatgcaggcgacaccaggcct-3′, 5′-atagaagcttctattgttccgcagcgtgca-3′. The fragment was cloned into the KpnI/HindIII site of pCold-GST (Hayashi and Kojima, 2008). The vector was transformed into OverExpress™ C43 (DE3) (Lucigen, WI, USA). The recombinant protein GST–AUR was purified with GSTrap HP (GE Healthcare, Buckinghamshire, UK). The protein was injected into a rabbit and the antiserum was subjected to affinity purification (Protein Purify Ltd, Gunma, Japan).

The anti-TOP guinea pig antibody was used at a dilution of 1∶1000 for immunoblotting, or at 1∶100 for immunofluorescence. Antibodies against Dynamin 1 (Dnm1), Dynamin 2 (Dnm2), Mda1, α-tubulin, CENH3 and mitochondrial porin were used as previously described (Nishida et al., 2003; Miyagishima et al., 2003; Nishida et al., 2005; Maruyama et al., 2007; Nishida et al., 2007; Fujiwara et al., 2009). Secondary antibodies used for immunofluorescence were, Alexa Fluor 488 or Alexa Fluor 555 goat anti-guinea pig, anti-mouse or anti-rabbit IgG, highly cross-adsorbed (Molecular Probes, Eugene, OR). Images were captured using a BX51 fluorescence microscope (Olympus, Tokyo, Japan), equipped with an XF37 narrow bandpass filter (Omega, Tokyo, Japan) and a C7780-10 three charge-coupled device (CCD) camera system (Hamamatsu Photonics, Shizuoka, Japan). Primary antibody reactions were performed for 1 hour at 4°C. Secondary antibody reactions were performed for 1 hour at 4°C. For the staining of phosphorylated protein, cells or isolated MD/PD machineries were stained with Pro-Q Diamond phosphoprotein gel stain (Molecular Probes, Eugene, OR) for 15 minutes at 4°C.

For immunoelectron microscopy (immuno-EM), primary reactions were performed for 1 hour at 4°C with guinea pig anti-TOP or rabbit anti-Mda1, diluted 1∶100 in Can Get Signal Immunostain Solution B (TOYOBO, Osaka, Japan), and labeled with gold particle-conjugated secondary antibody (British BioCell International, Cardiff, UK; 15-nm for TOP or 10-nm for Mda1) at a dilution of 1/20. After MD and PD machineries with outer membranes were incubated in organelle membrane dissolution buffer (OMD buffer; PBS containing 100 mM n-octyl-β-D-glucopyranoside, 6 mM sodium lauryl sulphate, 20 mM urea) for 2 minutes on ice, the lysate was negatively stained with 0.5% phosphotungstic acid (pH 7.0). The samples were then examined with an electron microscope (model JEM-1230; JEOL, Tokyo, Japan).

PCR reactions were performed with KOD FX (TOYOBO, Osaka, Japan) using the oligonucleotide primers, 5′-cttaaccgtactgatcgtact-3′ and 5′-tagtctaaactgagaacagcc-3′ and pBSHAb-T3′ as a template (Ohnuma et al., 2008), and 5′-tctcagtttagactactgcactcaaagtgagtgtccg-3′ and 5′-atcagtacggttaagtcatgtttgacagcttatcatc-3′ and pI050P-GFP as a template (Ohnuma et al., 2009). The resultant DNA fragments were combined and cloned to make pCPG (catalase promoter with sGFP) using the In-Fusion™ advantage PCR cloning kit (Clontech, CA, USA). The 5′-flanking region (1500 bp) and the open reading frame (ORF) of the TOP gene (2553 bp) was amplified using the oligonucleotide primers, 5′-ggcggccgctctagagaatggaaatcgcgcgcttctcc-3′ and 5′-tgggtaattaattaatggctcctggaaagagactcgttg-3′. The pCPG vector fragment was amplified using the oligonucleotide primers, 5′-ttaattaattacccatacgatgttcctgactatgcggg-3′ and 5′-tctagagcggccgccaccg-3′ and pCPG as a template. The resultant DNA fragments were combined and cloned to make pCPG-TOP-S using the In-Fusion™ advantage PCR cloning kit (Clontech, CA, USA). The antisense strand of the TOP ORF (2553 bp) was amplified using the oligonucleotide primers, 5′-aagtgcgcctgcgcatggctcctggaaagagactcgttg-3′ and 5′-tgggtaattaattaaatgattcgagacagg-3′. Using the oligonucleotide primers, 5′-ttaattaattacccatacgatgttcctgac-3′ and 5′-tgcgcaggcgcacttg-3′ and pCPG-TOP-S as a template, DNA fragments were PCR amplified, then combined and cloned to make pCPG-TOP-AS using the In-Fusion™ advantage PCR cloning kit (Clontech, CA, USA). pCPG-TOP-AS has the antisense strand of the TOP ORF instead of the sense strand of the TOP ORF. The resultant plasmids, pCPG as a control and pCPG-TOP-AS for antisense suppression of TOP, have a catalase promoter-sGFP fused sequence; therefore, transformed cells could be detected by green fluorescence under microscopic observation. Transformation of C. merolae cells was performed as described previously (Ohnuma et al., 2009).

We thank T. Shimada for technical help with MALDI-TOF analysis.