Human ASPM participates in spindle organisation, spindle orientation and cytokinesis

ASPM is a nuclear protein that relocates to the spindle pole matrix and central spindle MT minus ends during mitosis

To determine the intracellular distribution of human ASPM, rabbit polyclonal antibodies were raised against ASPM-specific peptide sequences (Figure 1A) and used to screen a panel of cell lines by immunofluorescence. Two antibodies against N-terminal peptides (216-1 and 217-2) and one antibody against a C-terminal peptide (279-3) produced virtually identical staining patterns (see Additional file 1, 2, 3) in a variety of human cell types including HeLa, primary human dermal fibroblasts (HDF), U2OS osteosarcoma cells, SH-SY5Y neuroblastoma cells and in COS-7 African green monkey kidney cells. Pre-incubation with the associated ASPM peptide negated the ASPM signal (data not shown). As previously published [37], we found ASPM expression to be predominantly concentrated in the nucleus during interphase (Figure 1B), however we did not find evidence of centrosomal distribution. Consistent with this finding, the ASPM sequence contains a number of potential nuclear localization sequences (Figure 1A). No specific association with microtubules or microtubule-organizing centres was observed prior to nuclear envelope breakdown. In agreement with previous reports, ASPM immunoreactivity shifted away from condensing DNA and onto spindle structures immediately after nuclear envelope breakdown, and from prometaphase to anaphase ASPM was seen almost exclusively associated with the mitotic spindle (Figure 1C) [8, 37, 38]. We localized mitotic ASPM immunoreactivity to pole-proximal regions of spindle microtubules, where it formed a ring around the spindle poles. This ring enclosed but did not extensively overlap with γ-tubulin immunoreactivity (Figure 1D), though it co-localised with Dynactin 1 (DCTN1) immunoreactivity in the pericentriolar matrix (PCM, data not shown). In anaphase cells the mitotic spindle has flattened and ASPM maintains its spindle pole localisation, was observed along the microtubules leading to the pole and a fraction of ASPM was found at the minus ends of central spindle microtubules (Figure 1E). Throughout telophase and cytokinesis ASPM was predominantly positioned in a narrow ring at the centre of the midbody (Figure 1F and see Additional file 4). This localisation was most visible when using the C-terminal antibody 279-3.

Immunoblot analysis using the 217-2 N- terminal anti-ASPM antibody identified a 410 kDa ASPM band in each of a panel of human and monkey cell lines (Figure 1G). This band corresponds to the expected size of the 3447 amino acid wild-type ASPM protein. Other investigators have reported that at least three predominant ASPM splice variants may exist [8]. In accordance with the reported literature, 217-2 would be predicted to recognise wildtype ASPM and the largest variant V1, however we saw no evidence for the existence of V1 in our whole cell extracts. We were unable to optimise antibodies 216-1 and 279-3 for Western blotting at endogenous levels of ASPM. However, these antibodies recognise recombinant forms of the ASPM protein expressed in and purified from bacteria (results not shown).

ASPM distribution is microtubule dependent

To examine the dependence of ASPM localization upon microtubules, we first treated HeLa cells with 5 μg/ml of the microtubule depolymerising drug nocodazole for 60 minutes. Under these conditions γ-tubulin immunostaining was maintained at spindle poles but ASPM 216-1 staining was lost, indicating that the mitotic distribution of ASPM is microtubule-dependent (Figure 2A). We next challenged HeLa cells with 10 μM taxol for 60 minutes, a treatment that induces the formation of multiple microtubule asters in mitotic cells. ASPM 216-1 immunostaining was observed at the centre of these asters (Figure 2B), suggesting a specific localization to microtubule minus ends. ASPM was not specifically associated with the anastral γ-tubulin foci also seen in taxol-treated cells (Figure 2C).

SiRNA mediated ASPM knockdown leads to repositioning of the mitotic spindle and cytokinesis dysfunction

To determine the effects of loss of ASPM function upon cell cycle progression we carried out ASPM RNA interference experiments. As a normal human NPC culture amenable to RNAi studies was not available we chose to perform these tests in the extensively characterised and readily transfectable U2OS osteosarcoma cell line using two individual ASPM siRNAs (ASP1 and ASP2) and a luciferase (GL3) control. A clear reduction in ASPM at the spindle poles of dividing cells was detected after transfection with ASP1 and ASP2 for 72 or 96 hours. By 72 hours post transfection with ASP1, ASPM (detected with the anti-ASPM 216-1 antibody at the spindle poles of mitotic cells was detectable in only 24% (6/25) of metaphase cells and completely knocked down in 76% of dividing cells, whereas ASP2 siRNA achieved a partial knockdown in 92% (23/25) metaphase cells (Figure 3A) and complete knock down in 8% of dividing cells. Comparative immunofluorescence analysis of average integral intensity measurements of immunofluorescently labelled centrosomal ASPM was performed for twenty-five metaphase U2OS cells treated with either ASP1, ASP2 or GL3 siRNAs. In comparison to the average ASPM spindle pole signal measured in the GL3 treated metaphases, a significant decrease of 84.3% in the average integral intensity of ASPM signal at the spindle pole was calculated after 72 hr treatment with ASP1, and 55.2% after treatment with ASP2 (Table 1). Immunofluorescence and Western blotting could not detect a change in overall expression level of ASPM in interphase cells (Figure 3B,C), suggesting that ASPM is stabilised when associated with the nuclear matrix and is more open to degradation after nuclear envelope breakdown. We therefore chose to investigate the phenotype associated with knock down of ASPM at the spindle poles of dividing cells.

A substantial reduction in proliferation of U2OS cells was observed after siRNA treatment in a Celltitre Blue cell viability assay (Figure 3D). Assessment of the proportions of 2N:4N DNA content by flow cytometry 48 hours and 72 hours after ASPM siRNA treatment distinguished 4N cells in G2/M from tetraploid cells that had failed cytokinesis by confirming the presence or absence of Cyclin B1. We identified an increase in the proportion of 2N cells in G1, a reduction of cells in G2/M and an increase in tetraploid cells in ASPM siRNA treated cells (Figure 3E). These data are consistent with both a decrease in the number of cells undergoing mitosis and a failure of cytokinesis. This observation was corroborated by microscopic examination of immunofluorescently labelled cells after 48 and 72 hours of ASP1 and ASP2 siRNA knockdown, which indicated a decrease in mitotic index and an increase in multinucleate cells and large mononucleate cells with multiple centrosomes consistent with a failure to complete cytokinesis (Figure 3F). The number of cells in cytokinesis, defined in this instance as daughter cells linked by narrow cytoplasmic bridges containing midbodies, was also decreased by ASPM siRNA.

To gain further insight into the ASPM knockdown phenotype, we carried out phase contrast time-lapse imaging of living U2OS cells 72 hours post-ASP1 or GL3 siRNA treatment. The results of four independent experiments were combined to produce the final data set. We monitored 126 cell divisions of GL3 treated U2OS cells and 94 cell divisions following ASP1 siRNA. Depletion of ASPM had a profound effect on U2OS cell division. Three major phenotypic anomalies were identified; an increase in the proportion of divisions with the mitotic spindle perpendicular to the plate surface, an increase in cytokinesis failure, and an increase in apoptosis.

Normal symmetrical division necessitates positioning of the mitotic spindle parallel to the surface of the imaging dish resulting in cleavage plane orientation perpendicular to the dish (Figure 4A, see Additional file 5). In 37% of ASP1 siRNA knockdowns (n = 36) random deviation in the position of the mitotic spindle was observed (Table 2) with a significant increase in spindle positioning approximately perpendicular to the surface of the dish (Figure 4A, see Additional file 6). This spindle alignment is reminiscent of that observed in asymmetric divisions in mouse progenitor cells of the neuroepithelium [35]. We therefore termed these divisions asymmetric. In comparison, only 10.3% (n = 13) of GL3 treated cells exhibited such asymmetric divisions.

A second consequence of ASPM depletion observed by time-lapse imaging was cytokinesis failure (Table 2), leading to the formation of multinucleated cells. Although we observed this phenomenon in a minor population of ASP1 siRNA-treated cells that underwent symmetrical division (6.4%, n = 6), the phenotype was far more prevalent in ASP1 treated cells that had undergone asymmetric divisions (28.7%, n = 27) (Figure 4A, see Additional file 6). In contrast, cytokinesis failure was observed in only 3.2% (n = 4) of symmetric cell divisions and 4% (n = 5) of asymmetric cell divisions in GL3 treated cells. In this culture system asymmetric cell division initially results in one daughter cell anchored to the culture dish with the second daughter cell positioned above the first (i.e., being extruded towards the observer see Additional file 6) and not attached to the substrate. We observed that in many instances the second daughter cell was able to subsequently make contact with the dish and successful cytokinesis occurred. Only 38% (5 of 13 cells) of asymmetrical divisions in GL3 control experiments failed cytokinesis compared with 75% (27 of 36 cells) of asymmetrical ASP1 treated cells. Therefore asymmetrical division per se does not fully account for increased cytokinesis failure in ASP1 siRNA treated U2OS cells.

Thirdly, a visible increase in cell death was observed in cells transfected with ASP1, 72 hours post transfection. Apoptosis was observed in mononucleate cells, binucleate cells (Figure 4A, see Additional file 7) and in cells exiting mitosis but exhibiting cytokinesis failure. To quantitate the extent of ASP1 siRNA induced apoptosis we performed an in situ activated caspase-3 assay on unfixed U2OS cells after 96 hours of siRNA treatment (Figure 4B). On average, ASP1 siRNA treated cells demonstrated a 2.7 fold increase in caspase-3 positive apoptotic cells compared to control GL3 siRNA cells (mean percentage of apoptotic GL3 siRNA cells in a total cell population = 7% (s.d. = 3.3%) compared to mean percentage of apoptotic ASP1 siRNA cells = 18.7% (s.d. = 6%) (Figure 4C).

It has previously been reported that ASPM is required for precise orientation of the mitotic spindle in specialised mouse NPC [35]. Our data demonstrates that ASPM function is a determinant of spindle position and division symmetry even in non-neuronal derived cells, suggesting that complete absence of ASPM function during human development might be expected to affect whole body development. Since this is not the observed phenotype in MCPH, the effect of ASPM mutation on the function of the ASPM protein expressed in MCPH patients was a logical progression of our studies.

An ASPM C-terminal splice donor site mutation induces novel splicing and reduces mitotic spindle pole association

ASPM contains a number of potential functional domains. Following the positional cloning of the ASPM gene, the initial genotype-phenotype correlation studies in MCPH patients did not reveal a clustered distribution of microcephaly-associated mutations in ASPM[5, 15] and there were therefore few clues to which of these domains were most important for ASPM function.

We have previously characterized a C-terminal ASPM mutation as a homozygous intronic mutation at IVS25 +1G > T (9984 +1 G > T), identified in individuals of a consanguineous family of Northern Pakistani origin [15]. This mutation lies downstream of the Armadillo repeat-like domain (Figure 5A). Splice site prediction programs forecast the outcome of this mutation to be the removal of the intron 25 splice donor site, resulting in translational extension of exon 25 by 29 amino acids before the incorporation of a stop signal. We predicted this would result in either the synthesis of a truncated ASPM protein of 395 kDa (3357 amino acids) that would not include the anti-ASPM 279-3 peptide sequence, or that it could lead to ASPM depletion due to mRNA degradation via the NMD pathway.

To determine the effect of the IVS25+1G > T mutation on ASPM mitotic expression and localisation, comparative immunofluorescence microscopy of dividing ASPM ^wt and ASPM ^{IVS25 +1G > T} fibroblasts was performed using our N-terminal 216-1 anti-ASPM antibody. Quantitative analysis of spindle pole staining revealed a significant reduction in ASPM levels at the PCM in ASPM ^{IVS25 +1G > T} fibroblasts (Figure 5B, C) (n = 80, average integral intensity ASPM ^wt = 16427.01 μm² (s.e.m. = 1166.3 μm²), average integral intensity ASPM ^{IVS25 +1G > T} = 2325.1 μm² (s.e.m. = 260.1 μm²), p = < 0.0001; Average area ASPM ^wt = 4.80 μm²(s.e.m. = 0.27 μm²), average area ASPM ^{IVS25 +1G > T t} = 1.28 μm² (s.e.m = 0.08 μm²), p = < 0.0001). Unexpectedly, analysis of ASPM ^{IVS25 +1G > T} fibroblasts immunofluorescently labelled with the C-terminal anti-ASPM 279-3 antibody also revealed weak but clearly visible spindle pole staining (Figure 5B), suggesting the IVS25+1G > T mutation did not induce a protein truncation.

To investigate the effect of the IVS25+1 G > T ASPM mutation on the ASPM protein, we carried out immunoblotting of cell lysates from control (ASPM ^wt ) and patient (ASPM ^{IVS25 +1G > T} ) fibroblasts with the 217-2 N-terminal anti-ASPM antibody. Expression of full length ASPM (410 kDa) in ASPM ^wt cell lines was observed and a band of approximately similar size and equivalent intensity was detected in ASPM ^{IVS25 +1G > T} protein extracts (Figure 5D). Thus the IVS25 +1 G > T mutation does not instigate NMD or protein instability in fibroblasts. Due to the large size of both the full length ASPM and the predicted IVS25+1G > T truncated product (395 kDa) we were unable to directly determine size differences in the proteins by immunoblotting.

To further explore the effect of the IVS+1G > T mutation, RT-PCR of ASPM exons 24-28 of ASPM ^wt and ASPM ^{IVS25 +1G > T} fibroblast mRNA was performed and the single PCR product obtained from each was sequenced. Wildtype ASPM sequence was observed in the ASPM ^wt fibroblast mRNA. As expected the G > T mutation at the exon 25 splice donor site invalidated the native splice site sequence in the ASPM ^{IVS25 +1G > T} mRNA. However, transcription did not continue for the expected 29 amino acids before the incorporation of a STOP codon. Unexpectedly, an in-frame cryptic splice donor site nine base pairs upstream was recognised and utilised (Figure 5E, F). The use of the upstream cryptic splice donor site would result in the production of an ASPM protein from which only a tripeptide motif (amino acids 3326-3328) is deleted (Figure 5E). Our immunofluorescence data therefore indicates that the loss of three amino acids in the C-terminus of ASPM is apparently sufficient to drastically reduce the ability of the protein to localise with the microtubules at the PCM at the spindle poles. The fact that fibroblasts from this MCPH patient have a significant decrease in the quantity of microtubule associated ASPM suggests the presence of a functional C-terminal domain in ASPM that is critical for normal ASPM function.

Expression of C-terminal fragments of ASPM inhibits spindle assembly and induces mitotic delay in HeLa cells

Our immunostaining data in patient cells indicates a function mediated by the extreme C-terminal region of ASPM is required for its localization to spindle poles. We therefore examined the consequences for mitotic cells of expressing C-terminal fragments of ASPM upon mitotic progression. Plasmids directing the expression of GFP fused to three different fragments of 301 (D1), 222 (D2) and 163 (D3) amino-acids from the C-terminal sequence of ASPM were created (Figure 6A) to determine whether ectopic expression of these fragments would disrupt the function of endogenous ASPM in a dominant-negative manner.

HeLa cells were transfected with the plasmids and observed 24 hours post transfection. Each construct led to similar abnormalities that included the generation of a large number of cells with morphological features consistent with an early mitotic arrest (Figure 6B, panels a-c) and in many cells membrane blebbing was also evident (Figure 6B, panels b-c). The DNA within these cells was highly condensed and occasionally fragmented, while spindle microtubule organisation was severely disrupted. Where visible, microtubule focusing and spindle formation appeared to be fundamentally impaired (Figure 6B, panel a, arrowheads). In most cells the density of microtubules was decreased (Figure 6B, panel b, arrowheads). γ-tubulin immunostaining invariably revealed the presence of two separate foci in these cells (Figure 6B, panel c, arrows), consistent with the hypothesis that these were cells displaying a disrupted mitosis. Endogenous ASPM immunoreactivity was dispersed (not shown). This is consistent with the observation of diminished microtubule density and our earlier finding that the ASPM spindle pole association is microtubule-dependent. It is also consistent with the interpretation that expression of these ASPM fragments directly inhibited the localisation of endogenous ASPM to spindle poles. Of control cells expressing GFP alone, 3% (n = 245) displayed an apoptotic or mitotic delay phenotype, in comparison with 33% (n = 135) of cells expressing ASPM-D1-GFP; 26% (n = 127) of cells expressing ASPM-D2-GFP; and 17% (n = 131) of cells expressing ASPM-D3-GFP. Similar observations were obtained in transfected COS-7 cells (not shown). We conclude that a high proportion of HeLa cells expressing ASPM C-terminal fragments, particularly those containing the whole of the armadillo repeat-like region (D1 and D2), are defective in assembling a functional spindle. This leads to mitotic arrest and eventual apoptosis.

Expression of ASPM COOH-terminal GFP-fusion proteins induces cytokinesis failure in COS-7 and HeLa cells

In addition to the severe mitotic defects seen in COS-7 and HeLa cells transfected with C-terminal fragments of ASPM, we also observed a subset of cells that appeared to have undergone mitosis but had failed to complete cytokinesis (Figure 6B, panels d-f). Defects ranged from cells containing large, highly lobular nuclei and peripheral microtubule bundles (Figure 6B, panel d, arrows) to cells containing discrete nuclei separated by aberrant microtubule arrays (Figure 6B, panel e, arrow), or structures resembling residual central spindles (Figure 6B, panel f, arrow). In neither case was there clear evidence for the presence of a cleavage furrow.

To exclude fixation artefacts we used co-expression of a GFP-tagged microtubule plus-end binding protein, EB3 [39, 40] as a means of observing centrosomes and the tips of growing microtubules and midbody structures in living COS-7 cells. Expression of low levels of EB3-GFP alone had no significant impact on cell division in transfected COS-7 cells (J. M. Askham, unpublished data, 41). The phenotypes of COS-7 cells co-expressing ASPM-D2-GFP and EB3-GFP confirmed our immunostaining observations (Figure 6C). Expression of ASPM-D2-GFP alone led to the formation of binucleated cells (Figure 6C panel a). Co-expression of EB3-GFP revealed that these cells contained central spindle structures or loose parallel microtubule arrays (Figure 6C, panel b-d). In some examples centrosomes were seen in close association with well-separated nuclei (Figure 6C, panel d), indicating that the primary defect in these cells was a failure in cytokinesis after the completion of chromosomal segregation. The observations from both fixed and living COS-7 cells were therefore consistent with a role for ASPM in organizing the central spindle microtubule array in mitotic cells from anaphase onwards.

Cell culture

HeLa and COS-7 cells were obtained from the European Collection of Cell Cultures. U2OS human osteosarcoma cells and SH-SY5Y neuroblastoma cells were obtained from the ATCC Cell Biology Collection (ATCC-LGC Promochem Partnership, South London, UK). Human neonatal dermal fibroblasts (ASPM ^wt ) were obtained from Genlantis (San Diego, California). Each cell line was maintained in accordance to manufacturer's guidelines.

The MCPH5 patient fibroblast cell culture (ASPM ^{IVS25+1G > T} ) was created from a forearm skin biopsy. The collection of biopsy sample was approved by the Ethical Committee of Leeds (East) Research Ethics Committee (REC reference number 05/Q1206/80) and informed consent was collected from the subject and parents. The skin biopsy was finely cut and homogenised to desegregate the cells prior to plating in Hams F10 medium containing L-glutamine and 25 mM HEPES (Invitrogen, Paisley, U.K.), 20% fetal calf serum (Invitrogen) and 1% penicillin/streptomycin (Invitrogen) and incubated in 5% CO₂ at 37°C

Antibodies and Cytotoxic reagents

Rabbit polyclonal antibodies against peptide sequences corresponding to aa363-386 (sequence code 216-1; KDNYGLNQDLESES), aa426-441 (sequence code 217-2; PEDWRKSEVSPRIPEC) and aa3443-3458 (sequence code 279-3; SRLKPDWVLRRDNMEE) of the human ASPM sequence were raised by Affiniti Research Products Limited, Exeter, UK. BLAST analysis revealed no significant homology to other published sequences. Antibody specificity in immunostaining was demonstrated by peptide pre-competition of immune serum and lack of specific staining from pre-immune serum (not shown). A rat anti α-tubulin antibody was obtained from Serotec Ltd, Oxford, UK. A mouse monoclonal antibody specific for γ-tubulin was obtained from the Sigma-Aldrich Company, Dorset, UK. Mouse monoclonal and rabbit polyclonal anti-GFP antibodies were obtained from Clontech (Saint-Germain-en-Laye, France). Secondary antibodies were Molecular Probes highly cross-absorbed rabbit, mouse or rat IgG-specific Alexa conjugates (Cambridge BioScience, Cambridge, UK). DAPI, nocodazole and taxol were obtained from the Sigma-Aldrich Company.

Immunofluorescence

Cells were cultured on glass coverslips. Transfections of GFP-constructs were performed using GeneJuice (Novagen) according to the manufacturer's instructions and coverslips were processed for immunocytochemistry 24 hours after transfection. Cells were fixed for 5 minutes in ice-cold methanol and labelled as previously described [48]. Immunofluorescence analysis was performed using a Leica TCS-SP confocal microscope as described previously [49] and a Zeiss Axiovert 200 inverted microscope coupled to an Orca II ER CCD camera controlled by AQM6 software (Kinetic Imaging, Wirral, UK) as described previously [50]. All confocal images presented here are maximum intensity projections. Figures were assembled and annotated using Adobe Photoshop CS2.

Immunoblotting

To perform immunoblotting of ASPM, cells were lysed in standard RIPA buffer, protein concentration determined and PAGE carried out using 3-8% Tris Acetate precast gradient gels (Invitrogen) in a Surelock electrophoresis tank (Invitrogen) on 25 μg protein. Subsequent protein transfer onto PVDF membrane (Invitrogen) was performed in the same system in accordance with the manufacturer's instructions.

siRNA knockdown of ASPM expression

Double-stranded RNA oligos that target ASPM mRNA were obtained from Dharmacon Inc Lafayette, CO, USA. siRNA ASPM1 targets sequence UGCCAUGGUGCAACUUGCU (nt 691-709) (synthetic oligos UGCCAUGGUGCAACUUGCUUU and AGCAAGUUGCACCAUGGCAUU) and siRNA ASPM2 targets GUGGUGAAGGUGACCUUUC (nt 2804-2822) (synthetic oligos GUGGUGAAGGUGACCUUUCUU and GAAAGGUCACCUUCACCACUU). A GL3 siRNA targeting the firefly luciferase sequence CUUACGCUGAGUACUUCGA was used as a control [51].

5 × 10⁵ U2OS cells per well of a six well plate containing a coverslip or 4 × 10⁵ U2OS cells per 35 mm plate for live cell imaging experiments were transfected 16 hours post plating with siRNA using Oligofectamine transfection reagent (Invitrogen). siRNA transfection mix was prepared for each plate as follows; 12 μl of 20 μM siRNA was added to 200 μl Optimem serum free medium (Invitrogen). 8 μl of Oligofectamine was added to 52 μl Optimem medium in a separate tube and incubated at room temperature for 10 minutes. The diluted transfection reagent was added to the tube containing the diluted siRNA and incubated for a further 20 minutes. 128 μl of Optimem was then added to give a final siRNA transfection mix volume of 0.4 ml. The growth medium was removed from the cells which were washed once with growth medium without serum or antibiotics. 0.6 ml growth medium without serum or antibiotics was added to the plate, followed by the 0.4 ml transfection mix. The transfection was incubated at 37°C for 4 hours before adding 1 ml growth medium containing 20% FCS and antibiotics. The cells were incubated for a total of 48 or 72 hours before analysis.

FACS analysis of DNA content

48 and 72 hours after siRNA treatment U2OS cells were released by trypsin digestion, pelleted, washed in PBS and resuspended in 1 ml of cold 70% ethanol then stored at -20°C. Prior to Flow Cytometric analysis the cells were washed in block buffer (PBS containing 1% BSA), permeabilised for 15 minutes in block buffer containing 0.25% Triton X-100 and then incubated for 1 hour in 100 μl of mouse anti cyclin B1 (BD-Pharmingen 554177 diluted 1:500). Cells were washed in block buffer then incubated for 1 hour with Alexa Fluor 488 anti-mouse IgG (Molecular probes A-11029, diluted 1:300). Cells were finally washed and then incubated at 37°C for 30 min in 0.5 ml PBS containing 3 μl of 6 mg/ml RNAse and 12.5 μl of 5 mg/ml propidium iodide. 10,000 single cell events per sample were analysed for DNA and cyclin B1 content to identify G1 (2N DNA content, low cyclin B1), G2/M (4N DNA content, high cyclin B1) and tetraploid cells (4N DNA content, low cyclin B1) using a Facscalibur Facscan (Becton Dickinson).

CaspGLOW™ Fluorescein Active Caspase-3 Assay

U2OS cells were treated with GL3 or ASP1 siRNA duplex for 96 hours prior to commencing Caspase-3 analysis. Detached and adherent cells were collected and washed with PBS. The CaspGLOW™ Fluorescein Active Caspase-3 assay (Medical and Biological Laboratories Ltd, Woburn, MA, USA) was performed on 3 × 10⁵ GL3 and ASP1 treated cells as per manufacturer's protocol. The cells were then resuspended in PBS containing 50 ug/ml propidium iodide, 100 ug/ml RNAse A and 0.1% (v/v) Triton X-100 and incubated at 37°C for one hour. Active caspase-3 was detected by flow cytometry using a BD LSRII (Becton Dickinson) capturing 10000 events per sample. Three independent experiments were performed

Time-lapse live cell imaging

Phase contrast time-lapse imaging experiments of live cells were performed on a Zeiss Axiovert 200 M inverted microscope incorporating an automated motorised stage, contained within a Perspex incubation chamber heated to a constant 37°C temperature (Solent Scientific, UK). Cells were cultured, transfected and imaged in 35 mm glass-bottomed culture dishes (Iwaki brand; Asahi Techno Glass Corporation, Japan, obtained from Barloworld Scientific, UK). Growth media was replaced with 2 ml media supplemented with 20 mM HEPES. Cells were observed under oil with an ×40 Plan Neofluar 1.3 aperture objective. Images were captured with a Hamamatsu Orca Camera, using Volocity 3 acquisition software (Improvision, Coventry, UK) over a period of 4 hours. Multiple positions were imaged at a rate of one image per minute. Live cell fluorescence imaging experiments were performed on a Zeiss Axiovert 200 inverted microscope contained within a Perspex incubation chamber heated to a constant 37°C temperature (Solent Scientific, UK). Imaging was performed as described previously [50]. Figures were assembled and annotated using Adobe Photoshop CS2.

Celltitre Blue cell viability assay and assessment of mitotic abnormalities in fixed immunolabelled siRNA treated cells

To perform cell viability assays on siRNA treated cells, U2OS cells were plated in 96-well plates at 10³ cells per well and grown at 37°C for 16 hours before siRNA treatment (as above). Wells were washed once with serum free medium and then 60 μl serum free medium and 40 μl siRNA transfection mix were added. The transfection was completed as described above. 72 hours after siRNA transfection, Celltitre Blue reagent (Promega) was added to wells following manufacturer's instructions. The plate was incubated at 37°C for a further 8 hours and the conversion of resazurin to fluorescent resorufin, an assay of metabolic rate, was determined using fluorimetry (560 nm excitation wavelength, 615 nm emission wavelength). In parallel experiments siRNA treated and control cell cultures were co-immunostained using combinations of antibodies that included α-tubulin, γ-tubulin and ASPM. DNA was counterstained using DAPI. At least 1000 total cells per sample were then assessed by direct microscopic observation to determine mitotic index, the number of cells in cytokinesis, and the proportion of cells with abnormal numbers of nuclei and centrosomes.

Comparative analysis of ASPM spindle pole expression in ASPM ^wt , ASPM ^{IVS25 +1G > T} cell lines and GL3, ASP1 and ASP2 treated U2OS cells

To compare relative quantities of ASPM at the spindle poles of ASPM ^wt and ASPM ^{IVS25 +1G > T} cells, cells were simultaneously immunofluorescently labelled using 216-1 anti-ASPM antibody. Images were captured on an Olympus BX61 upright microscope using ×100 oil UPlanFLN 1.30 aperture objective, F-View II monochromatic camera (Olympus Soft Imaging Solutions Ltd, Helperby, North Yorkshire, U.K.) and Cell^P Software (Olympus Soft Imaging Solutions Ltd), with a 1 second exposure time, ensuring the signal intensities visualised in the images were directly comparable. ASPM PCM signal intensity was measured as integral intensity (defined as the sum of all the intensities of a region of interest multiplied by the pixel area). To compare the integral integrity values of ASPM localised at the spindle pole in ASPM ^wt and ASPM ^{IVS25 +1G > T} cells a paired two tailed Student t-test was used. A P-value of less than 0.0001 was considered highly statistically significant.

To compare the integral integrity values of ASPM localised at the spindle pole in GL3, ASP1 and ASP2 treated metaphase U2OS cells captured with a 2 second exposure time, a Mann Whitney test was used. A P-value of less than 0.005 was considered statistically significant.

RNA preparation, reverse transcription and sequencing

RNA was obtained from ASPM ^wt and ASPM ^{IVS25 +1G > T} fibroblasts using Trizol extraction as per manufacturers instructions. cDNA was produced using Superscript reverse transcriptase (Stratagene Agilent Technologies, West Lothian, UK) as per manufacturers direction with oligo dT primers (Promega). ASPM exons 25-28 were amplified using forward (TCCGAAGTTGTAATCGCAGT) and reverse (CTTGCAGGGGATTTGTGATT) primers.

Plasmid construction

The plasmids pASPM-D1-GFP, pASPM-D2-GFP and pASPM-D3-GFP were constructed by cloning nt 9530-10434 (aa3177-3477), 9766-10434 (aa3256-3477) and 9943-10434 (aa3315-3477) of ASPM respectively in frame into the Bam HI and Xho I restriction sites in the GFP expression plasmid pJMA2-eGFP [49]. All constructs were fully sequenced before use. The construction and characterization of an EB3-GFP expression plasmid will be described elsewhere (J. M. Askham, manuscript in preparation).