K. Sakaguchi, Department of Applied Biological Science, Tokyo University of Science, 2641 Yamazaki, Noda, Chiba 278–8510, Japan Fax: +81 4 7123 9767 Tel: +81 4 7124 1501 (ext. 3409) E-mail: email@example.com
Sulfoquinovosylacylglycerols (SQAGs), in particular compounds with C18 fatty acid(s) on the glycerol moiety, may be clinically promising antitumor and/or immunosuppressive agents. They were found originally as inhibitors of mammalian DNA polymerases. However, SQAGs can arrest cultured mammalian cells not only at S phase but also at M phase, suggesting they have several molecular targets. A screen for candidate target molecules using a T7 phage display method identified an amino acid sequence. An homology search showed this to be a mammalian mitotic centromere-associated kinesin (MCAK), rather than a DNA polymerase. Analyses showed SQAGs bound to recombinant MCAK with a KD = 3.1 × 10−4 to 6.2 × 10−5m. An in vivo microtubule depolymerization assay, using EGFP-full length MCAK fusion constructs, indicated inhibition of the microtubule depolymerization activity of MCAK. From these results, we conclude that clinically promising SQAGs have at least two different molecular targets, DNA polymerases and MCAK. It should be stressed that inhibitors of MCAK have never been reported previously so that there is a major potential for clinical utility.
Several kinds of synthetic sulfoquinovosylacylglycerols (SQAGs) may be potent and clinically promising agents for cancer chemotherapy and immunosuppression [1,2]. However, the molecular targets of SQAGs are ambiguous. The aim of the present study was to identify molecular targets that could be of significance. Earlier, effects of SQAGs were found and reported independently by two laboratories screening directly for antitumor agents in vivo and for mammalian DNA polymerase inhibitors in vitro[3–5]. One molecular target of SQAGs is thus DNA polymerases  but there is a strong evidence for other targets [7–9].
SQAGs are found as natural compounds in higher plants , sea algae [4,5] and sea urchins . They have been reported to have a wide range of bioactivities: antiviral activity against human immunodeficiency virus (HIV-1) , P-selectin receptor inhibition , antitumor activity , tumor cell growth inhibition  and immunosuppressive activity . Sahara et al. showed that SQAGs effectively inhibit the growth of implanted human lung adenocarcinoma cells, A549, in nude mice . Moreover, Ohta et al. reported that a wide variety of cultured tumor cells were sensitive to SQAGs . It is very difficult to collect and purify SQAGs from natural sources but we have succeeded in the chemical synthesis of a number of forms. We have found α-anomers that possess potent antitumor activity but do not have many of the serious side-effects of standard cancer chemotherapeutics . In contrast, β-anomers did not show antitumorogenicity but were toxic to lymphocytes . The active SQAGs are 1-O-monoacyl-3-O-(α-d-sulfoquinovosyl)-glyceride with saturated or unsaturated fatty acids, and 1,2-O-diacyl-3-O-(α-d-sulfoquinovosyl)-glyceride and 1,2-O-diacyl-3-O-(β-d-sulfoquinovosyl)-glyceride, both with saturated fatty acids . The degree of inhibitory activity is greatly dependent upon the size of the fatty acid; SQAGs with fatty acid elements with fewer than 14 carbons do not show inhibitory effects in vitro or in vivo[10,11]. Our studies have identified possible discrepancies with regard to mechanistic aspects of SQAG activity in cancer cells. These compounds are considered to block replicative DNA synthesis by suppressing the activity of the DNA polymerases, thus arresting the cell cycle at S and consequently killing the cancer cells. However, aphidicolin, a well-established DNA polymerase inhibitor with cytotoxicity very similar to SQAGs, shows little bioactivity in vivo. As shown previously, moreover, SQAGs arrest cells not only at the S but also at M phase . These observations allow us to speculate that other molecular targets may be involved in vivo, possibly inducing cell death.
T7 phage display methods are powerful and high throughput tools for in vitro and in vivo identification of peptides or protein ligands. In this study, we used a T7 phage display method in combination with immobilized biotinylated SQAG prepared on an avidin solid phase [14–16]. A sequence was thereby identified that exhibited similarity with human mitotic centromere-associated kinesin (MCAK). Kinesins are molecules that convert chemical energy into physical reactions to perform functions such as vesicle transport, chromosome segregation, and organization of the mitotic spindle. Therefore, one of the other targets of the SQAGs is probably a MCAK. We show here that SQAGs suppress microtubule depolymerization by binding to MCAK. To our knowledge, this is the first report of an inhibitor to MCAK.
Screening for peptide sequences selectively binding to SQAG in the T7 phage random peptide library
We selected four representative SQAGs for binding analysis: 1-O-monoacyl-3-O-(α-d-sulfoquinovosyl)-glyceride with saturated [α-SQMG(18:0)] or unsaturated fatty acid [α-SQMG(18:1)]; 1,2-O-diacyl-3-O-(α-d-sulfoquinovosyl)-glyceride with saturated fatty acid [α-SQDG(18:0)]; and, 1,2-O-diacyl-3-O-(β-d-sulfoquinovosyl)-glyceride with saturated fatty acid [β-SQDG(18:0)](Fig. 1A). Although the distribution of α- and β-anomers in the body would be expected to differ , the levels of the cytotoxicity are similar when the size of the fatty acid is the same . Each chemically pure compound was synthesized in our laboratory. For screening peptides specifically binding to SQAGs, 1.5 × 108 p.f.u. per 30 µL of the T7 phage library expressed random peptide sequences was applied onto streptavidin-coated wells bearing an immobilized SQAG biotinylated derivative (Fig. 1B). We found that effective biopannning required a number of rounds of elution with 1.5 m NaI followed by washing with 0.1% Tween-20 in 100 mm Tris/HCl (pH 8.0). An illustrative example of the results of biopanning is shown in Fig. 2. For β-SQDG(18:0), the recovery rate of round 5 (i.e. the eluted fraction of 5th biopanning) was 7.7%, which was almost sixfold higher than those of rounds 1–4. The DNA sequences of 47 clones picked from round 5 were analyzed, and finally, an oligopeptide sequence was obtained as the clone which was highly concentrated. It was composed of 14 amino acids (NSRMRVRNATTYNS), and hereafter is called ‘clone-14’ for convenience.
When the binding titer of the phage ‘clone-14’ on the β-SQDG(18:0)-solid phase was compared to the unselected clone (Fig. 3), the recovery rate of the former was 5.1-fold higher. The ‘unselected clone’ which harbored five amino acids (NSNTR), was hardly concentrated in round 5 at all. The data indicate that ‘clone-14’ was effectively concentrated in the biopanning procedure, presumably due to selective binding to the β-SQDG(18:0) molecule. As indicated below, as with the other α-anomeric SQAGs used, α-SQMG(18:0), α-SQMG(18:1) and α-SQDG(18:0) also bind tightly and selectively to ‘clone-14’, binding must be unrelated to the anomeric structure (Table 1).
Table 1. SPR analysis of the binding of SQAGs to the immobilized peptide, MCAK184 on a CM5 sensor chip. A synthetic peptide and MCAK184 were coupled to the CM5 sensor chips. Binding analyses of SQAGs were performed in running buffer (Experimental procedures) at a flow rate of 20 µL·min−1 at 25 °C. biaevaluation 3.1 software was used to determine the kinetic parameters.
A homology search (fasta3) demonstrated that the amino acid sequence of ‘clone-14’ is similar to the ‘neck region’ of the rat, human and mouse mitotic centromere-associated kinesin (MCAK) (Fig. 4) [17–19]. Kinesin family proteins generally contain the motor domain in the N- or C-terminal of the primary sequence, and the position predicts the direction of walking on microtubules. MCAK belongs to the Kin I subfamily and its motor domain, unlike most kinesins, is in the interior of the protein. Moreover, the protein localizes at centromeres, performs roles in the depolymerization of microtubules, and affects chromosome segregation. The neck region is adjacent to the N-terminus of the motor domain in MCAK. From a study using a truncated form, the neck region appears to be essential for microtubule depolymerization activity. Our previous data indicated that SQAGs arrest cultured cells not only at the S phase but also at M phase, the place in which microtubule depolymerization occurs . Therefore, SQAGs may inhibit microtubule depolymerization activity and thereby induce cell death. For this reason, we focused on the molecular interactions between SQAGs and the recombinant MCAK.
Kinetic parameters via surface plasmon resonance (SPR) analysis of binding between SQAG and MCAK
The full length MCAK protein is not soluble and is found in inclusion bodies. Therefore binding between SQAGs and a truncated version (MCAK184) was tested using a Biosensor BIAcore instrument. The MCAK184 contains the neck region and the Kinesin motor domain (Fig. 5). Three or four different concentrations of each of the four SQAG [1–4] derivatives shown in Fig. 1 were employed for analyses of the bindings to CM5 sensor chip conjugated 14aa or MCAK184. The dissociation constants KD (m) were determined using the biaevaluation 3.1 software (Table 1). Values with 14aa were in the range of KD = 1.3 × 10−5 to 8.7 × 10−7m, and for MCAK184 were KD = 3.1 × 10−4 to 6.2 × 10−5m.
SQAG inhibits the depolymerization activity of MCAK in vitro
To test the possibility that the interaction of SQAGs and MCAK184 inhibited depolymerization of MTs, we performed an in vitro depolymerization assay in the same manner as reported previously [20–22]. The truncated MCAK constructed as a His6-tagged truncated form (P184-G593; MCAK184), containing the neck and motor domains (Fig. 5A), was purified to near homogeneity (Fig. 5B). Depolymerization of the tubulin polymer could be detected in SDS/PAGE as an abundance of a band of tubulin molecules released into the supernatant. The in vitro depolymerization reactions contained 120 nm MCAK184 and 1500 nm of taxol-stabilized microtubules (taxol-stabilized MTs). MCAK184 depolymerized MTs in vitro, when incubated at 24 °C for 30 min, but only in the presence of 1 mm ATP (Fig. 6A). The presence of the ATPase inhibitor AMPPNP abolished depolymerization activity (Fig. 6A). Figure 6B shows the effects of 19.6 µmα-SQMG(18:1) and 3.2 µmβ-SQDG(18:0) on the depolymerization reaction. The concentrations of the SQAGs used were chosen from the minimum inhibitory concentration (MIC) with MCAK184. In this case, the reaction mixture contained 2% dimethylsulfoxide (DMSO), because of the solubility of the SQAGs. DMSO had no effect on the reaction (Fig. 6B, upper panel). Both α-SQMG(18:1) and β-SQDG(18:0) clearly inhibited the depolymerization (Fig. 6B, middle and lower panels). α-SQMG(18:0) and α-SQDG(18:0) tended also showed the same inhibition pattern (data not shown). Thus, at least under in vitro conditions, SQAGs inhibit the depolymerization activity of MCAK by selective binding. The tightness of binding may determine the degree of inhibition.
SQAGs inhibit the depolymerization activity of MCAK in vivo
To elucidate the MT-depolymerizing effects of SQAGs in vivo, the fusion constructs of EGFP-full length MCAK were transfected into cultured cells. After fixation, the cells were stained for tubulin and DAPI (Fig. 7) and digital images were acquired using a cooled CCD camera. Loss of microtubule polymers was observed in controls not treated with α-SQMG(18:1) (Fig. 7Ab, unfilled arrow), indicating that the polymers are rapidly depolymerized. However, in the presence of α-SQMG(18:1), the polymers were not depolymerized (Fig. 7Ae).
Figure 7B shows numbers of stained cells at various concentrations of α-SQMG(18:1). The data are from three independent experiments. The numbers of the cells showing depolymerization of tubulin decreased in a dose-dependent manner with increase in α-SQMG(18:1). Similar effects were exhibited by the other SQAGs (data not shown).
In the present study, we have shown, using a T7 phage display method, that mammalian mitotic centromere-associated kinesin (MCAK) is a molecular target of SQAGs. The SQAGs inhibit the MCAK function and are likely to bind to its ‘neck region’. As this M-type kinesin is localized at centromeres and depolymerizes microtubules from their ends, an important feature of remodeling during mitosis [20–23], it is conceivable that the anticancer effects of SQAGs are dependent on the inhibition of MCAK.
Although the SQAGs were found as inhibitors of mammalian DNA polymerases [3–5, 24], the impact on these enzymes appears too weak to explain their in vivo anticancer activity and their weak cytotoxicity. We reported previously that SQAGs can not only arrest cells at S phase but also at M phase ; thus the two different cell cycle phases may be impacted simultaneously.
SQAGs can be separated roughly into two groups according to the number of fatty acid molecules: diacyl-forms (SQDGs) and monoacyl-forms (SQMGs). Both are sulfonic analogs of d-glucose bound with glycerol and fatty acids. Our present results showed that the various derivatives of SQAG may strongly inhibit MCAK activity. This inhibition may be independent of their anomeric forms, as it is the case for their DNA polymerase inhibitor. Chemical synthesis of SQMG/SQDG derivatives produces both α- and β-anomers. The β-anomer is not present in nature. As described previously, α-anomers of synthetic SQMG/SQDG derivatives could be potent antitumor agents without severe side-effects. In mice exposed to these agents, the immunosuppressive effect was minor and the main visceral organs showed no histological evidence of toxicity . On the other hand, the β-anomer may be potent immunosuppressive agents with toxic effects on lymphocytes . Therefore, synthetic SQMG/SQDG, chemically composed only of carbohydrate glycerol and fatty acids, could be ideal cancer-chemotherapeutic and/or immunosuppressive agents that could be applied clinically for longer periods. The reason for tissue-specific toxicity dependent on the different configuration can be explained by the inhibition of neither DNA polymerases nor MCAK, pointing to the possibility of further molecular targets.
As DNA polymerases are essential for DNA replication and repair, their inhibition will induce cell cycle arrest at the S phase. The MIC ranges for DNA polymerases in vitro were low at 1–50 µm[6,11], while cytotoxicity was evident at 50–100 µm. Inhibition of MCAK activity occurred at 0.8–20 µm. Therefore, cell death in vitro may occur as a result of synergistic actions. SQAGs are also known to act against inflammation , respiration of spermatozoa [26–28], HIV-RT (human immunodeficiency virus-reverse transcriptase) [4,5,7], AIDS virus , the P-selectin receptor  and α-glucosidase . With the exception of the last two, the in vivo molecular targets for these effects have yet to be elucidated. Interestingly, the binding analysis of SQAGs and MCAK184 showed that the kinetic constant (KD) for the interaction between α-SQMG(18:1) and MCAK184 was lowest recorded (6.2 × 10−5m) (Table 1). Of the SQAGs used here, α-SQMG(18:1) is the strongest anticancer agent , suggesting that the tightness of binding to MCAK is important for in vivo bioactivity. As described previously, the inhibition of DNA polymerases occurs by tight binding to molecular pockets on their surfaces. The degree of inhibition depends on the KD (m) between the SQAG and the enzyme.
Although there are many drugs that bind tubulin directly (such as paclitaxel or nocodazole, which over- or understabilize microtubules, respectively), a drug that targets the M-type kinesin has never been reported. Thus the SQAGs may be of particular significance, not only with regard to their clinical applications, but also for analysis of the functions of MCAK.
SQAGs and a biotinylated derivative of SQAG were synthesized in our laboratory (Fig. 1) [3–5,14].
Construction of a T7 phage library from Drosophila melanogaster
Random primers, 5′-methylated dCTP, T4 DNA polymerase, T4 ligase, EcoRI/HindIII linkers, EcoRI, HindIII, T7Select10–3b vector, and T7 packaging extracts were purchased from Novagen (Darmstadt, Germany). Construction of the phage library was carried out according to the manufacturer's instructions. In brief, aliquots (80 µg) of total RNA, extracted from D. melanogaster Kc cells, (provided by M Yamaguchi, Kyoto Institute of Technology, Japan) were used to construct a cDNA library. Total RNA was treated with Oligotex-dT30 super (Takara, Shiga, Japan) to produce poly(A)+ RNA suitable for random primed cDNA synthesis. cDNA synthesis was performed using 4 µg of poly(A)+ RNA. 5′-Methylated dCTP was then incorporated into both strands, without extraction or precipitation between the first and second strand synthesis. The cDNA was then treated with T4 DNA polymerase to generate flush ends and ligated with directional EcoRI/HindIII linkers. Following linker ligation, the cDNA was digested sequentially with EcoRI and HindIII, then inserted into EcoRI/HindIII digested T7Select10–3b vector arms. The cDNA was cloned into the EcoRI/HindIII site of the T7 phage 10–3b vector and packaged into phage . The titer of this library was 1.6 × 1010 p.f.u.·mL−1.
T7 phage clone biopanning procedure and DNA sequence analysis
A biotinylated derivative of SQAG was immobilized on a Streptavidin-coated ELISA plate (Nalge Nunc International, Wiesbaden, Germany) overnight at 4 °C. Unbound SQAG was removed by washing three times with 150 µL Tris buffer (100 mm Tris/HCl, pH 8.0) and plates were blocked with 200 µL of Tris buffer containing 3% (w/v) BSA for 1 h. The plates were washed three times with 200 µL of Tris buffer and then incubated, for 3 h with gentle rotation, with a T7 phage library composed of cDNA fragment inserts from D. melanogaster. Unbound phage was removed by washing three times with 0.1% (v/v) Tween-20 in Tris buffer. Bound phage was eluted from each plate by first an overnight incubation at 4 °C with 200 µL of 1.5 m NaI, and then four washes with 100 µL of 1.5 m NaI. The supernatant (total 600 µL) from both steps was collected and regarded as the eluted fraction. An aliquot (10 µL) was used to determine the titer of detached phage at each round of selection. The remainder was amplified by the plate lysate amplification method  for a new round of selection in the same manner as described above.
Following five rounds of selection, 47 plaques were arbitrarily picked up from LB plates and each dissolved in phage extraction buffer (20 mm Tris/HCl, pH 8.0, 100 mm NaCl, 6 mm MgSO4). In order to disrupt the phages, the extract was heated at 65 °C for 10 min. Phage DNA was then amplified by PCR, using T7 SelectUP and T7 SelectDOWN primers (T7Select Cloning kit, Novagen). PCR products were cloned in the pGEM-T vector (Promega, Madison, WI, USA) and sequenced using a DNA Sequencer 4200S (Aloka, Tokyo, Japan). From these sequence results, the amino acid sequence displayed on the T7 phage capsid was determined. A homology search (fasta3) demonstrated that the amino acid sequence of clone-14 is similar to the neck region of the MCAK. Some parameters were changed: Database, SwissProt; Expectation upper value, 50; Matrix, BL50; Number of alignments, 50.
Comparisons of affinity for SQAG with selected and unselected T7 phage single clones
The affinity of the candidate clone (positive) for SQAG was compared with that of an unselected clone (negative). The negative clone showed low selectivity with biopanning. Both single clone phages were amplified for liquid lysate amplification  and adjusted to a titer of 1.0 × 1013 p.f.u.·mL−1. One hundred microliters of each single phage suspension (i.e. 1012 p.f.u.) was applied onto SQAG immobilized plates. Washing and eluting proceeded as described above for biopanning. The titer of the eluted fraction was determined.
Construction of recombinant human MCAK
MCAK cDNA was derived from a human peripheral blood cDNA library by PCR (forward primer: 5′-ATGGCCATGGACTCGTCGCT-3′, reverse primer; 5′-TCACTGGGGCCGTTTCTTGC-3′). The neck and motor domains of MCAK cDNA (550–1770), conjugated with NdeI and XhoI restriction sites, were cloned into the pET21a expression vector (Novagen). EGFP-full length MCAK was made by the cloning of XhoI-BamHI MCAK cDNA fragment (forward primer: 5′-CTCGAGATGGCCATGGACTCGTCG-3′, reverse primer: 5′-GGATCCTCACTGGGGCCGTTTCTT-3′) into the pEGFP-C3 vector (BD Biosciences, Tokyo, Japan).
His6-tagged MCAK184 protein preparation
MCAK184 protein was overexpressed, purified for SPR analysis, and an in vitro MT depolymerization assay was conducted. Protein expression was performed by transforming the construct into BL21 (DE3)-pLysS (Novagen) and growing these bacteria in 1 L of Luria–Bertani medium containing 50 µg·mL−1 of kanamycin, 100 µg·mL−1 chloramphenicol. Cells were grown and treated with 1 mm of isopropyl thio-β-d-galactoside. After 3 h, they were harvested by centrifugation at 3000 g for 15 min. Cell pellets (3.5 g) were resuspended in 30 mL of ice-cold column binding buffer (20 mm sodium phosphate, pH 7.4, 0.5 m NaCl, 35 mm imidazole) and sonicated. Cell lysates were centrifuged at 27000 g for 20 min and the soluble protein fraction was collected as a crude extract and loaded onto a 5 mL HisTrap HP column (Amersham Biosciences, Foster City, CA, USA) of the FPLC system (ÄKTA 1 explorer, Amersham Biosciences) with a flow rate of 1 mL·min−1. The column was washed firstly with 100 mL binding buffer and then washed with 20 mL of buffer (20 mm sodium phosphate, pH 7.4, 0.5 m NaCl, 65 mm imidazole). Finally, MCAK184 was eluted with 100 mL of eluting buffer (20 mm sodium phosphate, pH 7.4, 0.5 m NaCl, 270 mm imidazole). For surface plasmon resonance (SPR) analysis, the eluted MCAK184 protein was dialyzed against HBS/EP buffer [10 mm Hepes, pH 7.4, 150 mm NaCl, 3 mm EDTA, 0.005% (v/v) Tween-20]. For the in vitro MT depolymerization assay, the eluted MCAK184 protein was dialyzed against sodium phosphate buffer (50 mm sodium phosphate, pH 7.0, 150 mm NaCl).
Surface plasmon resonance (SPR) analysis
The binding characteristics of SQAGs and a synthetic peptide ‘NSRMRVRNATTYNS’ (ANYGEN, Gwang-ju, Korea), MCAK184, was analyzed using a Biosensor 3000 instrument (BIAcore AB, Uppsala, Sweden) with CM5 research grade sensor chips (BIAcore). The synthetic peptide (332 µg·mL−1, 170 µL) in coupling buffer (10 mm sodium carbonate/sodium hydrogen carbonate, pH 8.5) was injected over a CM5 sensor chip at a 10 µL·min−1 of flow rate to capture the peptide on the carboxymethyl dextran matrix of the chip by using amine coupling at 25 °C. The surface was activated by injecting a solution containing 0.2 mN-ethyl-N′-dimethylaminopropyl carbodiimide (EDC) and 50 mmN-hydroxysuccimide (NHS) for 14 min. The peptide was injected and the surface was then blocked by injecting 1 m ethanolamine at pH 8.5 for 14 min. This reaction immobilized about 1500 resonance units (RU) of the peptide. When MCAK184 (332 µg·mL−1, 170 µL) in coupling buffer (10 mm acetic acid/sodium acetate, pH 4) was injected, about 2700 RU were immobilized. Binding analysis of SQAGs was performed in running buffer [10 mm Hepes, pH 7.4, 150 mm NaCl, 3 mm EDTA, 0.005% (v/v) Tween-20, 8% (v/v) DMSO] at a flow rate of 20 µL·min−1 at 25 °C. To measure the binding specificity and kinetics for 14 amino acids (aa), various SQAGs were injected for 120 s [α-SQMG(18:0): 22.8, 91, 364, 728 µm; α-SQMG(18:1): 0.2, 1.14, 2.28, 5.46 mm; α-SQDG(18:0): 60, 118, 236, 472 µm; β-SQDG(18:0): 7.4, 17.6, 23.5, 29.4 µm]. To measure MCAK184, various SQAGs were injected for 120 s [α-SQMG(18:0): 91, 364, 728 µm; α-SQMG(18:1): 204, 319, 364 µm; α-SQDG(18:0): 118, 235, 470 μm; β-SQDG(18:0): 70.6, 76.4, 82.3 µm]. Association and dissociation were each measured for 120 s at 20 µL·min−1. biaevaluation 3.1 software (BIAcore) was used to determine the kinetic parameters.
In vitro microtubule depolymerization assay
A microtubule depolymerization assay using polymerized, taxol-stabilized tubulin from bovine cytoskeleton (Denver, CO, USA), was performed as described previously [20–22]. For the assay shown in Fig. 6A, 120 nm MCAK184 in 20 µL of column eluting buffer (250 mm imidazole, pH 7.0, 300 mm KCl, 0.2 mm MgCl2, 0.01 mm Mg-ATP, 1 mm dithiothreitol and 20% glycerol) was mixed with 1500 nm taxol-stabilized microtubules in 80 µL of BRB80 (80 mm Pipes, pH 6.8, 1 mm EGTA, and 1 mm MgCl2), 12.5 µm taxol, 1 mm dithiothreitol, and 1.25 mm Mg-ATP/1.25 mm Mg-adenyl-5′-yl imidodisphosphate incubated at 24 °C for 30 min, and then centrifuged at 223 000 g for 15 min. For the assay shown in Fig. 6B, there was 2% (final concentration) DMSO in all samples. Supernatants and pellets were assayed for the presence of tubulin on Coomassie-stained SDS/polyacrylamide gels.
Cell transfection and immunofluorescence
CHO-K1 cells (Japan Health Sciences Foundation, Tokyo, Japan) were grown in Ham's F12 medium with 10% (v/v) fetal bovine serum. Transfection of pEGFP-C3-full length MCAK was performed with Lipofectamine™2000 (Invitrogen, Carlsbad, CA, USA). After transfection, cells were cultured for 24 h and various concentrations of SQAGs were administered. Cells were exposed for 24 h, fixed with 1% paraformaldehyde in cold methanol for 10 min, then incubated for 1 h with a mouse anti-(α-tubulin DM1A IgG) Ig (Sigma-Aldrich, St Louis, MO, USA) at 1 : 1000 dilution and a rhodamine-conjugated anti-mouse Ig (Chemicon International, Temecula, CA, USA) at 1 : 100 dilution in NaCl/Pi, 0.1% (v/v) Triton X-100 and 1% (w/v) BSA for 1 h. Finally they were washed with NaCl/Pi and mounted in mounting medium [NaCl/Pi, 4′,6-diamidino-2-phenylindole (DAPI), 10% (v/v) glycerol] for analysis under a microscope (Axioskop 2 plus, Zeiss, Tokyo, Japan). Digital images were acquired with a cooled CCD camera (AxioCam HRm, ZEISS) controlled by axiovision 3.0 software (Zeiss).