PLK1-mediated phosphorylation of WDR62/MCPH2 ensures proper mitotic spindle orientation

Abstract

Primary microcephaly (MCPH) is an autosomal recessive disorder characterized by congenital reduction of head circumference. Here, we identified compound heterozygous mutations c.731 C > T (p.Ser 244 Leu) and c.2413 G > T (p.Glu 805 X) in the WDR62/MCPH2 gene, which encodes the mitotic centrosomal protein WDR62, in two siblings in a Japanese family with microcephaly using whole-exome sequencing. However, the molecular and cellular pathology of microcephaly caused by WDR62/MCPH2 mutation remains unclear. To clarify the physiological role of WDR62, we used the CRISPR/Cas9 system and single-stranded oligonucleotides as a point-mutation-targeting donor to generate human cell lines with knock-in of WDR62/MCPH2 c.731 C > T (p.Ser 244 Leu) missense mutation. In normal metaphase, the mitotic spindle forms parallel to the substratum to ensure symmetric cell division, while WDR62/MCPH2-mutated cells exhibited a randomized spindle orientation caused by the impaired astral microtubule assembly. It was shown that a mitotic kinase, Polo-like kinase 1 (PLK1), is required for the maintenance of spindle orientation through astral microtubule development. In this study, we demonstrated that WDR62 is a PLK1 substrate that is phosphorylated at Ser 897, and that this phosphorylation at the spindle poles promotes astral microtubule assembly to stabilize spindle orientation. Our findings provide insights into the role of the PLK1–WDR62 pathway in the maintenance of proper spindle orientation.

Introduction

Primary microcephaly (MCPH) is an autosomal and recessive neurodevelopmental disorder characterized by a remarkable reduction of occipitofrontal head circumference (OFC) at birth to at least two to three standard deviations below the mean (1,2). To date, 13 genetic loci causing MCPH have been reported (1). The majority of MCPH gene products localize to centrosomes, which are nonmembranous organelles that consist of two perpendicular centrioles surrounded by an amorphous pericentriolar matrix. Centrosomes function as the microtubule-organizing centre (MTOC) throughout the cell cycle in mammalian cells to control diverse physiological events including cell migration, cell shape formation, and cell division (3). During mitosis, centrosomes form a bipolar spindle pole and assemble the microtubule network: kinetochore microtubules (MTs) connect to chromosomal kinetochores; interpolar and central spindle MTs locate the cleavage furrow upon cytokinesis; and astral MTs anchor the mitotic spindle to the cell cortex (4). Impaired centrosome function during mitosis induces chromosomal segregation errors, growth arrest, and apoptosis (3). Thus, the aetiological mechanisms of primary microcephaly are suggested to be attributable to defects in centrosome functions.

Mitotic spindle orientation is essential for cell fate determination, tissue morphogenesis, and organ size control (5). Proper spindle orientation depends on a physical connection between the cell cortex and astral MTs emanating from the centrosomes. In most cells, spindle orientation determines the position of the cell division plane to control whether a cell undergoes symmetric or asymmetric cell division. During human embryogenesis, the central nervous system develops through a series of symmetric and asymmetric cell divisions. At embryogenesis, the neuroepithelial stem cells (NESCs) with spindles parallel to the apical plane divide symmetrically to increase the population of NESCs. At the onset of neurogenesis, radial glial cells derived from NESCs form vertical spindles to divide asymmetrically, thereby producing the committed precursor cells of differentiated neurons (5). Indeed, several mutations of MCPH genes have been implicated in the control of spindle orientation (6,7). However, the molecular mechanism by which mutations of MCPH genes in human cells result in instability of spindle orientation remains poorly understood.

In this study, we identified compound heterozygous mutations of the WDR62/MCPH2 gene in a Japanese family with microcephaly using whole-exome sequencing. WDR62, which harbours 12 WD40 repeat-containing domains required for protein–protein interactions, localizes to mitotic centrosomes to assemble the MTs composed of spindles. Previous reports demonstrated that lymphoblastoid cells derived from patients with these mutations show mitotic spindle formation defects caused by insufficient MT assembly (8–10). In WDR62 mutant mice, premature differentiation of neurons induces impaired cortical development accompanied by dysregulation of the orientation of spindles in NESCs (11–13). Interestingly, it was shown that the fly orthologue of WDR62 is epistatically required for the centrosomal recruitment of the mitotic kinase Polo-like kinase 1 (PLK1) in order to control the mitotic spindle orientation in neuroblasts (14). In mammalian cells, PLK1 is known to play a critical role in stabilizing spindle orientation through astral MT nucleation (15). WDR62-associated biological links between spindle orientation control and the genetic interaction with PLK1 led us to explore how WDR62 and PLK1 cooperate to orchestrate spindle orientation in human cells. Here, we report that PLK1 phosphorylates WDR62 at S897 to develop the astral MTs and to stabilize spindle orientation parallel to the substratum.

Results

Clinical report of two siblings in a Japanese family with microcephaly

Patient II-2, a boy, was born at 38 weeks of gestation to healthy and nonconsanguineous parents (Supplementary Material, Fig. S1A). The pregnancy was unremarkable. An older brother was healthy without malformations. Birth weight was 2, 905 g (−0.3 SD), height 49.4 cm (+0.2 SD), and head circumference 30.9 cm (−1.5 SD). By the age of 20 months, he was able to stand and walk. At the age of 2 years 8 months, he was 87.9 cm in height (−0.9 SD) and had a head circumference of 44.8 cm (−2.7 SD). At the age of 4 years 11 months, significant developmental and growth delays persisted; his height was 102.2 cm (−0.9 SD) and his head circumference was 45.8 cm (−2.7 SD). He had microcephaly with mental retardation, but not seizures. G-banded chromosomes were 46, XY.

Patient II-3, a female sibling of patient II-2, was born at the gestational stage of 37 weeks 4 days, weighing 2,140 g (−2.1 SD) and measuring 43.6 cm (−2.2 SD) in height, with a head circumference of 28 cm (−2.8 SD) (Supplementary Material, Fig. S1A). The pregnancy was basically normal. However, ultrasound scan showed a foetus with a significantly small head. At the age of 2 months, she continued to exhibit severe developmental and growth retardation, with height of 50.8 cm (−3.2 SD) and head circumference of 31.5 cm (−4.4 SD). She did not have significant seizures despite the presence of spikes in the electroencephalogram at the age of 10 months. At the age of 11 months, she measured 66.3 cm (−2.8 SD) in height and 38.4 cm (−3.9 SD) in head circumference. Upon G-banding analysis, her karyotype was revealed to be 46, XX. Array CGH analysis did not detect significant copy number variations. The parents did not show microcephaly. Head circumference of mother I-1 was 54.5 cm (−0.4 SD) and that of father I-2 was 56.5 cm (+ 0.5 SD). Taken together, the clinical information suggested that the two siblings presented with autosomal recessive microcephaly (MCPH).

Whole-exome sequencing identified compound heterozygous mutations in the WDR62/MCPH2 gene

To identify a causative gene underlying the microcephaly in the two patients, we performed whole-exome sequencing of genomic DNA from both affected siblings and their parents (subjects I-1 and I-2). Average coverage for the exomes was more than 100×, and 96% of consensus coding sequence (CCDS) exons among all exomes were covered at more than 10× (Supplementary Material, Table S1). We narrowed down the candidate variants based on filtering criteria consisting of the public and internal databases (dbSNP, 1000 Genomes, and ∼300 internal exomes), genomic position, function, and zygosity. A total of 279 variants were common to subjects II-2 and II-3. Based on the linkage analysis, we identified the compound heterozygous variants c.731 C > T, p.Ser 244 Leu and c.2413 G > T, p.Glu 805X in the WDR62/MCPH2 gene (NM_001083961, OMIM 613583), which encodes a centrosomal protein with WD40 repeat-containing domains (Fig. 1A). The p.Ser 244 Leu variant was predicted to be ‘probably damaging’ by SIFT and Polyphen2. Sanger sequencing confirmed that subjects I-1 and I-2 were both heterozygous carriers of the WDR62/MCPH2 c.731 C > T and c.2413 G > T variants (Fig. 1B).

Figure 1.

Compound heterozygous mutations of the WDR62/MCPH2 gene encoding a mitotic centrosomal protein. (A) Schematic of WDR62 protein structure and WDR62 mutations in patients II-2 and II-3, as determined by whole-exome sequencing. The probands were a compound heterozygote with c.731 C > T, p.Ser 244 Leu in exon 7, and c.2413 G > T, p.Glu 805 X in exon 20. (B) Sanger sequencing confirmed the compound heterozygous mutations in this family. (C) RT-PCR analyses of exons 6–22 of WDR62 cDNAs in LCLs from the patients, the parents and a normal individual in an unrelated family. The patients had no transcripts of the WDR62 gene. The HPRT transcript was amplified in both samples as a positive control. (D) Western blotting analysis showing loss of WDR62 in the patients. The level of WDR62 was normalized by β-tubulin. (E) Immunostaining with anti-WDR62 (red) and anti-α-tubulin (green) antibodies in mitotic spindles of LCLs from patient II-3, mother I-1, father I-2 and a normal individual in an unrelated family. DNA was stained with DAPI (blue). Scale bar: 10 μm.

Next, we investigated the expression level of WDR62 mRNA in the cells of the patients and parents. RT-PCR analysis of exons 6–22 of WDR62 cDNA in the patient’s cells detected no transcripts, while the transcripts in the cells from both parents were significantly reduced compared with that of unrelated normal cells (Fig. 1C). Treatment of the patient’s cells with cycloheximide, a nonsense-mediated decay inhibitor (16,17), increased the level of WDR62 transcripts (Supplementary Material, Fig. S1B). We isolated the WDR62 cDNA from the patient’s cells treated with cycloheximide to determine the sequence of each clone. Of the 25 cDNA clones, 24 were from the nonsense mutation allele and 1 was from the missense allele (Supplementary Material, Fig. S1C). These findings imply that the WDR62 c.2413 G > T, p.Glu 805X variant leads to nonsense-mediated decay and the WDR62 c.731 C > T, p.Ser 244 Leu variant mediates an unknown mechanism, which reduce the level of WDR62 transcripts in patient cells. Western blotting and immunostaining analyses also revealed that lymphoblastoid cell lines (LCLs) from the patients had no WDR62 protein (Fig. 1D and E).

Consistent with previous reports on WDR62/MCPH2-mutated primary microcephaly (8–10), the patients’ cells showed impaired microtubule development of the mitotic spindle (Fig. 1E). LCLs from the parents showed the reduced amount of WDR62 protein compared with that of normal cells, but the microtubule development of spindle in the parent cells developed normally (Fig. 1D and E). We therefore concluded that the WDR62/MCPH2 gene is the most likely candidate for the causative gene underlying the autosomal recessive microcephaly in this family.

Generation of human cultured cells with biallelic knock-in of WDR62/MCPH2 c.731 C > T, p.Ser 244 leu mutation using genome editing technology

To confirm that the WDR62 c.731 C > T, p.Ser 244 Leu variant is indeed a causative mutation in the patients, we attempted to introduce this variant into normal cultured human cells. We constructed an expression vector driving both Cas9 protein and single guide RNA (sgRNA) to introduce DNA double-strand breaks into WDR62 exon 7 harbouring c.731 C > T, p.Ser 244 Leu (Fig. 2A). The expression vector carried the Cas9 gene and a puromycin resistance gene separated by a 2 A peptide sequence, allowing expression of the discrete protein products from a single open reading frame. As a targeting donor, we designed single-stranded oligonucleotides (ssODN) carrying the WDR62 c.731 C > T variant, putative silent CRISPR/Cas9-blocking mutations in the sgRNA sequence, and a silent BamHI sequence for checking ssODN knock-in easily (Fig. 2A and B). For genome editing experiments, the human colon cancer cell line HCT116 was used because it has two copies of the WDR62 allele and relatively high activity of homologous recombination for the knock-in of a targeting donor. We transfected both the Cas9-2 A-Puro expression vector and the ssODN into HCT116 cells. After transient puromycin selection for 48 h post-transfection, 75 colonies were picked up and their genotypes were analysed by BamHI digestion and direct sequencing of the PCR amplicon of the target region (Fig. 2B–D). Of these, five clones were shown to be monoallelically and two to be biallelically ssODN-targeted (WDR62^S244/S244L cells). We also generated a WDR62-null HCT1116 cell line (WDR62^−/− cells) by the biallelic insertion of part of the Cas9-2 A-Puro expression vector into WDR62 exon 20 (Supplementary Material, Fig. S2). RT-PCR, western blotting, and immunostaining analyses all demonstrated no WDR62 products in the WDR62^S244/S244L cells and WDR62^−/− cells (Fig. 2E–G). These results suggested that WDR62 c.731 C > T, p.Ser 244 Leu mutation pathogenetically influences the expression level of WDR62 product.

Figure 2.

Generation of WDR62^S244L/S244L HCT116 cells using the CRISPR/Cas9 system and ssODN. (A) WDR62 exon 7 sequencing alignment in an individual with c.731 C > T, p.Ser 244 Leu (purple-shaded square) with silent BamHI and CRISPR/Cas9 blocking mutations (grey-shaded squares). Orange arrowhead indicates the site of DNA double-strand breakage by the CRISPR/Cas9 system. (B) Restriction enzymatic map of WDR62 exon 7 targeting site. (C) Electrophoresis image of BamHI-digested PCR products of WDR62 exon 7 in the WDR62^+/+ and WDR62^S244L/S244L cells. (D) Sanger sequencing indicating confirmation of the biallelic introduction of WDR62 c.731 C> T, p.Ser 244 Leu and a silent BamHI mutation into HCT116 cells. (E) Loss of WDR62 mRNA in the WDR62^−/− and WDR62^S244L/S244L cells examined by RT-PCR. The HPRT transcript was amplified in all samples as a positive control. (F) Western blot analysis showing the depletion of WDR62 in the WDR62^−/− and WDR62^S244L/S244L cells. The level of WDR62 was normalized by β-tubulin. (G) Projective confocal images of the mitotic spindles in WDR62^+/+, WDR62^−/−, and WDR62^S244L/S244L cells immunostained with anti-WDR62 (red) and anti-α-tubulin (green) antibodies. DNA was stained with DAPI (blue). Scale bar: 10 μm.

Loss of the WDR62/MCPH2 gene causes uncontrolled spindle orientation in human cells

To explore the centrosomal function of WDR62 associated with the aetiological mechanism of primary microcephaly more precisely, we examined the numbers of centrosomes and chromosomes, as well as cell cycle progression, in WDR62^-/- and WDR62^S244L/S244L cells. They showed no significant changes compared with those in WDR62^+/+ cells, suggesting that WDR62 is dispensable for proper chromosome segregation and cell proliferation (Supplementary Material, Fig. S3). Most of the genes associated with primary microcephaly are known to be involved in the control of mitotic spindle orientation during neural development (3). Immunostaining analysis using anti-pericentrin and α-tubulin antibodies revealed that confocal z-section planes obtained from WDR62^+/+ cells contained two signals of spindle poles, indicating that WDR62^+/+ cells assemble a bipolar spindle oriented parallel to the substratum (Fig. 3A). In contrast, WDR62^S244L/S244L cells had a bipolar spindle misoriented relative to the substratum, such that spindle poles were detected in different confocal z-section planes (Fig. 3A). The quantitative analysis of spindle orientation also demonstrated that WDR62^S244L/S244L cells exhibited significantly larger spindle angles than WDR62^+/+ cells (Fig. 3B and C). These findings suggest that WDR62 mediates astral MT assembly to stabilize spindle orientation parallel to the substratum in human cells.

Figure 3.

WDR62 is required for proper spindle orientation. (A) Uncontrolled spindle orientation in WDR62^S244L/S244L cells. Immunostaining of the mitotic spindles in WDR62^+/+ and WDR62^S244L/S244L cells with anti-pericentrin (red) and anti-α-tubulin (green) antibodies. DAPI was used to detect DNA (blue). Representative images of z sections (0.3 μm per stack) with maximum centrosome intensity in metaphase cells are shown. Scale bar: 10 μm. (B) Diagram of mitotic spindle angle calculated by 1/tan (the distance between centrosomes along the z axis/the distance between centrosomes in the xy plane). (C) Quantification of (A) showing a significant increase of spindle angle in WDR62^S244L/S244L cells (*P < 0.05: t-test, n = 3: >50 cells per experiment). (D) Projective confocal images of the mitotic spindles in WDR62^+/+ and WDR62^S244L/S244L cells immunostained with anti-CAK5Rap2 (red) and anti-α-tubulin (green) antibodies. DNA was stained with DAPI (blue). Scale bar: 10 μm. The intensities of the spindle (I_spindle) and the astral and spindle (I_total) were measured with confocal microscope software. (E) Relative astral MT intensity (I_{asral, rel}) was calculated as follows: [(I_total −  I_spindle)/I_spindle]. i = z-stack position. (F) Quantification of (D) demonstrating a significant reduction of astral MT formation in WDR62^S244L/S244L cells (*P < 0.05: t-test, n = 3: >50 cells per experiment).

Since proper spindle orientation is required for the interaction of astral MTs with the cell cortex, we next examined whether astral MT development is impaired in WDR62^S244L/S244L cells. Projection of the confocal z-section images stained with anti-α-tubulin antibody revealed the significant reduction of astral MTs in WDR62^S244L/S244L cells (Fig. 3D–F). To track the MT plus-end tips of spindle and astral MTs, we stained cells with anti-MT plus-end-binding protein EB1 antibody (Supplementary Material, Fig. S4A). In WDR62^+/+ cells, the signals of EB1 were detected at the plus-ends of spindle MTs and astral MTs that contacted the cell cortex. However, in WDR62^S244L/S244L cells, EB1 was recruited at the plus-end tips of spindle MTs, but not near the cell cortex (Supplementary Material, Fig. S4A and B), implying that the physical interaction of astral MTs with the cell cortex was impaired. To evaluate whether the genetic background of HCT116 cells contributes to the phenotype of disrupted spindle orientation, we generated WDR62^S244L/S244L HEK293T cells using CRISPR/Cas9 system and ssODN. WDR62^S244L/S244L HEK293T cells also showed uncontrolled spindle orientation with reduced astral MTs, suggesting that the missense mutation of WDR62 affects cell division axis in the different genetic backgrounds (Supplementary Material, Fig. S5). These results imply that WDR62 at spindle poles promotes astral MT assembly to regulate correct spindle orientation.

PLK1-mediated WDR62 phosphorylation assembles astral microtubules to control spindle orientation

The mitotic kinase PLK1 is known to be a master regulator of spindle orientation in most animal cells. Since it was recently reported that WDR62 mediates microtubule assembly to recruit PLK1 to the centrosomes in Drosophila neural stem cells (neuroblasts) (14), we examined whether WDR62-dependent PLK1 localization to spindle poles is conserved in human cells. Immunostaining analysis using anti-phosphorylated PLK1 (T210) antibody, which precisely detects the activation of PLK1 since no signals were detected in WDR62^+/+ cells after treatment with BI2536, a PLK1 inhibitor, revealed that PLK1 localizes to the spindle poles and is activated in WDR62^S244L/S244L cells as well as WDR62^+/+ cells (Fig. 4A). These results indicated that the centrosomal targeting of PLK1 is independent of WDR62 in human cells. On the other hand, immunoprecipitation analysis demonstrated that PLK1 physically interacts with WDR62 in a polo-box domain (PBD)-dependent manner (Fig. 4B). We thus considered the possibility that WDR62 is not involved upstream of PLK1 but is a PLK1 substrate for controlling spindle orientation.

Figure 4.

WDR62 is dispensable for PLK1 activation at spindle poles, but they physically interact with each other. (A) WDR62^+/+ and WDR62^S244L/S244L cells were immunostained with anti-phosphor-T210-PLK1 (red) and anti-α-tubulin (green) antibodies. DNA was stained with DAPI (blue). In the presence of 20 nM BI2536 for 4 h, the phosphor-T210-PLK1 signals at the spindle poles were rarely detected. Scale bar: 10 μm. (B) Full-length or indicated mutant proteins of FLAG-tagged mouse PLK1 (mPLK1) and AcGFP1-tagged WDR62 were coexpressed in HEK293T cells and then immunoprecipitated from whole-cell lysates using the anti-FLAG antibody. AcGFP1-tagged WDR62 and FLAG-tagged mPLK1 fragments in the IP fractions and inputs were detected by western blotting.

To verify this, we attempted to identify the residue on WDR62 phosphorylated by PLK1 linked to the regulation of spindle orientation. WDR62 contains four conserved motifs for PLK1 phosphorylation (D/E-X-S/T-X_1-3-X-D/E) (18). To confirm whether any of these PLK1 motifs play a role in controlling spindle orientation, we constructed AcGFP1-tagged alanine-substituted mutants (T575A, S897A, S987A, and S1038A), which were expected to perturb the spindle orientation parallel to the substratum. We introduced them into WDR62^−/− cells and measured the spindle angles. The T575A, S987A, and S1038A mutants showed similar activity to wild-type WDR62, whereas the S897A WDR62 mutant did not lead to the restoration of proper spindle orientation (Fig. 5A and B). These results suggested that the S897 residue on WDR62 is involved in the PLK1-dependent maintenance of spindle orientation. To confirm that this residue is indeed the site of PLK1 phosphorylation, we raised anti-phospho-WDR62 (S897) polyclonal antibody and performed an in vitro kinase assay with His-tagged PLK1 and glutathione S-transferase (GST)-tagged WDR62. The antibody specifically recognized the wild type of WDR62 phosphorylated at S897 by PLK1 but not the WDR62 S897A mutant (Fig. 5C). Furthermore, immunostaining analysis revealed that phospho-WDR62 (S897) signals were detected at the spindle poles in WDR62^+/+ cells but not in WDR62^+/+ cells treated with BI2536 or in WDR62^S244L/S244L cells (Fig. 5D).

Figure 5.

PLK1-mediated phosphorylation of WDR62 at Ser 897 occurs at mitotic spindle poles. (A) WDR62^−/− cells were transfected with AcGFP1-tagged WDR62 or AcGFP1-tagged WDR62 with alanine substitutions at PLK1 phosphorylation consensus sites, and then immunostained with anti-GFP (green) and anti-pericentrin (red) antibodies. DNA was stained with DAPI (blue). The scale bar represents 10 μm. (B) Quantification of (A) showing that the S897A mutation of WDR62 significantly reduces the effect of rescuing uncontrolled spindle angle in WDR62^−/− cells (*P < 0.05: t-test, n = 3: >50 cells per experiment). (C) PLK1-mediated phosphorylation of WDR62 at Ser897 in vitro was probed with anti-phosphor-WDR62 (S897) antibody. The loading levels of His-PLK1 and GST-tagged C-terminal fragment (806–1160 a.a.) of WDR62 were also detected by western blotting. (D) WDR62^+/+ and WDR62^S244L/S244L cells were immunostained with anti-phosphor-S897-WDR62 (red) and anti-α-tubulin (green) antibodies. DNA was stained with DAPI (blue). The phosphor-S897-WDR62 signals at the spindle poles in WDR62^+/+ cells treated with 20 nM BI2536 for 4 h and WDR62^S244L/S244L cells were rarely detected. Scale bar: 10 μm.

To clarify the effect of WDR62 Ser897 phosphorylation on the control of spindle orientation, we introduced a WDR62 phosphomimetic mutant (S897E) into WDR62^+/+ cells treated with BI2536. Exogenous WDR62 S897E mutant corrected spindle misorientation caused by PLK1 inhibition compared with that in the wild type of WDR62 (Fig. 6A and B). In addition, the phosphomimetic WDR62 mutant, but not wild-type WDR62, also restored the impaired astral MT formation in BI2536-treated WDR62^+/+ cells (Fig. 6C and D). These findings suggested that the phosphorylation of WDR62 is required for the PLK1-dependent establishment of spindle orientation. We thus concluded that PLK1 at spindle poles phosphorylates WDR62 at Ser 897 to assemble astral MTs, thereby ensuring the proper spindle orientation in human cells (Fig. 6E).

Figure 6.

PLK1-mediated phosphorylation of WDR62 at Ser 897 is required for the control of mitotic spindle orientation. (A) WDR62^+/+ cells were transfected with AcGFP1-tagged WDR62 or PLK1-phosphorylation mimic WDR62 mutant (S897E) and cultured in the presence of 20 nM BI2536 for 4 h before immunostaining with anti-GFP (green) and anti-pericentrin (red) antibodies. DNA was stained with DAPI (blue). The scale bar represents 10 μm. (B) Quantification of (A) showing that the phosphor-mimic S897E WDR62 mutant significantly restored dysregulation of the spindle angle even in the absence of BI2536 (*P < 0.05: t-test, n = 3: >50 cells per experiment). (C) Projective confocal images of the mitotic spindles in the cells transfected with AcGFP1-tagged WDR62 or the S897E mutant. They were immunostained with anti-GFP (green) and anti-α-tubulin (red) antibodies. DNA was stained with DAPI (blue). Scale bar: 10 μm. (D) Quantification of (C) indicating that the phosphor-mimic S897E WDR62 mutant significantly promoted astral MT formation even in the absence of BI2536 (*P < 0.05: t-test, n = 3: >50 cells per experiment). (E) Model for the PLK1–WDR62 pathway in the control of mitotic spindle orientation. At mitotic spindle poles, WDR62 phosphorylated at Ser 897 by PLK1 enhances astral MT formation to stabilize the spindle orientation parallel to the substratum.

Discussion

Here, we identify unique compound heterozygous variants of the WDR62 gene in a Japanese family with microcephaly as the causative mutations, because WDR62 is the second most common cause of primary microcephaly and the variants identified were predicted to be damaging by annotation tools. However, candidate variants extracted by applying a forward genetic approach are not always predicted correctly in rare genetic disorders. To overcome this problem, a reverse genetic approach using genome editing technology in human cultured cells is useful. If the nucleotide variant is indeed the disease-causing mutation, its introduction into normal cultured human cells would result in the abnormal cellular phenotype of the disorder. Previously, we developed a TALEN-mediated two-step single-base-pair editing technology in human cultured cells, and demonstrated that a single-nucleotide substitution upstream of BUB1B, which encodes the mitotic checkpoint molecule BubR1, is a causative mutation for a rare autosomal recessive disorder characterized by a predisposition to cancer, severe microcephaly, and conditions on the ciliopathy disease spectrum (19,20). However, it was labour-intensive because it required multiple sets of TALENs and two rounds of drug selection. We therefore designed a single-step framework allowing knock-in of specific mutations into human cultured cells using Cas9-2 A-puromycin resistance gene and ssODN. Using this approach, we isolated two HCT116 cell clones with biallelic WDR62 S244L mutations and five clones with a monoallelic mutation out of 75 clones of the puromycin-resistant cells. However, all of the five clones with a monoallelic mutation contained an insertion or deletion in the second allele of the WDR62 locus, possibly caused by Cas9-mediated nonhomologous end joining. These results suggested that it was difficult to generate precise heterozygous cell clones rather than homozygous cell clones. Therefore, an alternative genome editing strategy is required for cellular modelling of a heterozygous carrier or a patient with a dominant genetic disorder. Recently, a method termed CORRECT (consecutive re-guide or re-Cas steps to erase CRISPR/Cas-blocked targets) was reported (18). This method enables the efficient generation of heterozygous human cultured cells, although it needs two genome editing steps (21). Therefore, further work aimed at improving this genome editing technology in human cultured cells should be performed to establish a rapid and precise method for identifying any genetic mutation.

PLK1-dependent astral MT formation is indispensable for correct spindle orientation (15). Here, we show that WDR62 is a PLK1 substrate at mitotic spindles and promotes astral MT formation to ensure the proper spindle orientation. A series of studies on the control of spindle orientation have revealed several direct substrates of PLK1 including LRRK1 kinase, NDR1 kinase, MISP, and Dishevelled (15,22–24). To our knowledge, this study is the first to provide evidence that PLK1 directly controls the MCPH gene product. Interestingly, Drosophila WDR62 acts epistatically upstream of PLK1 to maintain asymmetric centrosome maturation in neuroblasts despite the physical connection between WDR62 and PLK1 (14), suggesting that the PLK1–WDR62 axes involved in the development of the central nervous system are related but not strictly conserved between vertebrates and insects. It has also been implied that WDR62 is required to directly or indirectly stabilize MTs. The spindle association and function of WDR62 during mitosis coincide with its elevated phosphorylation status. Centrosome-associated Aurora kinase A phosphorylates WDR62 at S49 and T50 to promote the centrosomal targeting of WDR62 for shaping the mitotic spindle through MT stabilization (25). In contrast, c-Jun N-terminal kinase (JNK) phosphorylates WDR62 at T1053 to dissociate WDR62 from spindle MTs, thereby inhibiting MT stabilization (26,27). PLK1 inhibition by BI2536 does not alter the centrosomal localization of WDR62 (Fig. 6A and B), suggesting that PLK1-mediated phosphorylation of WDR62 at Ser897 is involved in the activation of MTOC activity at spindle poles for producing astral MTs, independently of the subcellular targeting of WDR62. Although further investigations are required to clarify how WDR62 promotes the formation of astral MTs, our findings demonstrate the PLK1-mediated phosphorylation of WDR62 functions as a molecular switch for the formation of astral MTs.

A predominant model for the aetiology of primary microcephaly is that defects in spindle orientation cause altered symmetric/asymmetric cell divisions of NESCs during the neural developmental process (5). However, this model is controversial because in vivo studies in mouse embryos have shown that randomized spindle orientation does not necessarily induce a small brain (28). Thus, cellular processes involved in the aetiology of a small brain remain unclear. Recent studies have revealed that WDR62 knockout mice show not only uncontrolled spindle orientation but also the failure of kinetochore-spindle MT attachment during neurogenesis (29). WDR62 interacts with the kinetochore-associated Aurora B kinase to regulate chromosome segregation and mitotic progression (12). Moreover, the kinetochore-associated genes CASC5 and CENPE are causative of MCPH4 and MCPH13, respectively (1). These results suggest that the PLK1–WDR62 axis at spindle poles for maintaining MTOC activity might control kinetochore–spindle MT interaction, although WDR62^S244L/S244L cells did not show significant aneuploidy (Supplementary Material, Fig. S2C) in this study. WDR62 S244L mutation is not directly related to the Plk1-catalyzed phosphorylation of WDR62 on S897 for the development of astral MTs, while the mutation reduces the levels of mRNA and protein to impair the proper spindle orientation. Further study will be necessary to understand the molecular basis of primary microcephaly.

In conclusion, we demonstrate that genomic editing technology in human cultured cells is useful for the precise diagnosis of causative mutations, and that the PLK1–WDR62 pathway is essential for the astral MT formation underlying proper spindle orientation.

Materials and Methods

Whole-exome analysis

Exonic DNA was captured from genomic DNA extracted from the peripheral lymphocytes of subjects I-1, I-2, II-2, and II-3 using SeqCap EZ Human Exome Library v2.0 (Roche NimbleGen). Sequencing was performed with 100-bp paired-end reads on a HiSeq2000 (Illumina). We used BWA (http://bio-bwa.sourceforge.net/; date last accessed September 1, 2017) for alignment and mapping, Samtools (http://samtools.sourceforge.net/; date last accessed September 1, 2017) and Picard (http://broadinstitute.github.io/picard/; date last accessed September 1, 2017) for SAM/BAM handling, GATK (http://www.broadinstitute.org/gatk/; date last accessed September 1, 2017) and Samtools for variant calls, and Annovar (http://annovar.openbioinformatics.org/; date last accessed September 1, 2017) for annotation, as described previously (30). Functional predictions due to amino acid changes were estimated using PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2; date last accessed September 1, 2017), SIFT (http://sift.bii.a-star.edu.sg/; date last accessed September 1, 2017), and Mutation Taster (http://www.mutationtaster.org/index.html; date last accessed September 1, 2017). Control exome sequences were obtained from Japanese patients undergoing exome analysis for diseases other than primary microcephaly. All reported genomic coordinates were in GRCh37/hg19. PCR amplification followed by Sanger sequencing with an Applied Biosystems 3130 sequencer (ThermoFisher) was used to validate mutations identified by whole-exome sequencing and to examine the segregation of variants within the family. The primer sequences were as follows: forward: 5′-TATCATCTGGTCATGCAGGG-3′, reverse: 5′-AACCTGTGTGGTCAATGCTG-3′ for exon 7; and forward: 5′-ACCACCAGCCCATTTGC-3′, reverse: 5′-AACAAGGTCATGGGCAAAAC-3′ for exon 20 of the WDR62 gene.

Cell cultures

Informed consent was obtained from the members of the Japanese family with microcephaly prior to this study. LCLs were established from peripheral blood lymphocytes of patient II-3. We used LCLs from an apparently healthy female in an unrelated family as a control. LCLs were cultured in RPMI 1640 with 20% foetal bovine serum at 37 °C with CO₂. HEK293T and HCT116 cells were maintained in DMEM supplemented with 10% foetal bovine serum at 37 °C with CO₂. The transfection of expression plasmid vectors into cells was performed using Lipofectamine LTX (ThermoFisher) or nucleofection with Nucleofector^TM 2 b device (Lonza), in accordance with the manufacturers’ protocols. The cell cycle profile was analysed using Muse™ Cell Analyzer (Millipore).

Antibodies

The primary antibodies used were: rabbit anti-WDR62 polyclonal antibody pAb (Bethyl Laboratories; A301-560 A); rabbit anti-CDK5Rap2 pAb (Merck Millipore; 06-1398); rabbit anti-pericentrin pAb (Bethyl Laboratories; A301-348 A); rat anti-α-tubulin mAb (Novus; NB600-506); mouse anti-γ-tubulin mAb (Sigma Aldrich; T6557); mouse anti-β-tubulin mAb (Sigma Aldrich; T8328); mouse anti-GAPDH mAb (Santa Cruz Biotechnology; sc-32233); rabbit anti-GFP pAb (MBL; 598); rabbit anti-PLK1 pAb (Santa Cruz Biotechnology; sc-5585); rabbit anti-phospho-T210 PLK1 pAb (BioLegend; 618602); mouse anti-EB1 antibody (BD Transduction Laboratories; 610535); and mouse anti-DYKDDDDK (FLAG)-tagged mAb (Wako; 018-22381).

Anti-phosphorylated S897-WDR62 pAbs were raised in two rabbits using keyhole limpet hemocyanin (KLH)-conjugated ESLEN(pS)ILDSL peptide. Antibodies were affinity-purified on phosphorylated- and/or nonphosphorylated-epitope-bound columns.

Plasmids

3 × FLAG-tagged human PLK1 and its mutants were as described previously (31). We constructed an AcGFP1-tagged or GST-tagged human WDR62 plasmid by PCR and standard cloning steps. We also used site-directed mutagenesis to insert mutations into WDR62. For the construction of an expression vector of both spCas9 and sgRNA targeting WDR62 gene exon 7 or exon 20, a pair of annealed oligodeoxynucleotides designed on the target site (WDR62 exon 7: 5′-TGTTGTTGTGCAGCTCGCCC-3′, WDR62 exon 20: 5′-GAGAGCAAACAGAGGATGATC-3′) with overhangs of the BbsI restriction enzyme site were inserted into the pX459-U6-Chimeric_BB-CBh-hSpCas9-2 A-Puro V2.0 plasmid (Addgene; #62988). The successful integration of mutations or oligodeoxynucleotides into each plasmid vector was verified by Sanger sequencing using an Applied Biosystems 3130 sequencer (ThermoFisher).

Generation of WDR62-edited HCT116 cells using genome editing technology

200 pmol 100-mer ssODN carrying the WDR62 c.731 C > T, p.S244L variant and a silent BamHI site (5′-CTGTGTCTCCAGGTGACGAGCACAGTGCCCCTTGTAGGGCGCTtGGGGATCCTTGGGGAGCTGCACAACAACATCTTCTGTGGTGTGGCCTGCGGTCGGG-3′) and 1 μg of the pX459 plasmid vector for editing of exon 7 of the WDR62 gene were cotransfected into 1 × 10⁶ HCT116 cells and HEK293T cells with Kit V (Lonza) and program D-032 (HCT116) or Q-001(HEK293T) (Nucleofector^TM 2 b device; Lonza), in accordance with the manufacturer’s protocol. For the generation of WDR62^−/− cells, we introduced the pX459 plasmid vector for editing of exon 20 of the WDR62 gene into the HCT116 cells under the same conditions. At 48 h after transfection, the transfected cells were subjected to selection for 48 h using 100 μg/ml puromycin (Nacalai Tesque). The drug-resistant cell colonies were then picked up 14 days later. These colonies were divided into two aliquots: one was transferred into a well of a 96-well plate for clonal expansion, while the other was lysed and used for PCR with a pair of primers for the amplification of WDR62 exon 7 or exon 20. For the screening of WDR62^S244L/S244L cells, PCR fragments were digested with BamHI and then run on 2.0% agarose gel. The BamHI-sensitive products were analysed by direct-sequence genotyping. For the screening of WDR62^−/− cells, the PCR products were run on a 2.0% agarose gel and subjected to direct-sequence genotyping.

RT-PCR

Total RNA was extracted from HCT116 cells and LCLs using TRIzol reagent (Life Technologies), in accordance with the manufacturer’s protocol. LCLs were treated with 100 μg/ml cycloheximide (Sigma Aldrich) for 2 h at 37 °C before RNA extraction. First-strand cDNAs were generated with random hexamers using M-MLV reverse transcriptase (Life Technologies). To examine the expression of WDR62 mRNA, primers in exon 6 (5′-AAAGACATCGTAGTGGCCTCCAACA-3′) and exon 22 (5′-TGGTAGCTGCGCTTGGCGTGGAAG-3′) were designed. HPRT was assayed as an internal control (forward: 5′-GAAGAGCTATTGTAATGACC-3′; reverse: 5′-GCGACCTTGACCATCTTTG-3′). PCR amplification was carried out with KOD-Fx neo DNA polymerase (Toyobo).

Immunoprecipitation and western blot analyses

Cells were transfected with plasmid DNA and cultured in serum-free D-MEM for 24 h, after which they were lysed in lysis buffer (0.5% Triton X-100, 150 mM NaCl, 20 mM Tris-HCl pH 7.5, 1 mM EDTA, 0.5 mM PMSF, 2 mg/ml pepstatin A, 10 mg/ml leupeptin, 5 mg/ml aprotinin). The lysates were sheared with a 21-gauge needle, incubated on ice for 15 min, and clarified by centrifugation at 20, 817 g for 15 min at 4 °C. The supernatants were precleared with Protein A/G-conjugated magnetic beads (Merck Millipore) and incubated with anti-FLAG antibody for 2 h at 4 °C with constant rotation. Protein A/G-conjugated magnetic beads were then added to the lysates and the mixtures were rotated for a further 16 h at 4 °C. The magnetic beads were washed three times with wash buffer (1% Nonidet P-40, 0.1% SDS, 0.5% deoxycholate, 150 mM NaCl, 50 mM Tris-HCl pH 7.5, 1 mM EDTA, 0.5 mM PMSF, 2 mg/ml pepstatin A, 10 mg/ml leupeptin, 5 mg/ml aprotinin) before elution with sample buffer. The immunoprecipitated proteins were analysed by 10% SDS-PAGE and transferred to PVDF membranes for western blot analyses, as described previously (19).

Immunofluorescence microscopy

Cells grown on coverslips treated with collagen (Matsunami Glass) were fixed in 4% paraformaldehyde at room temperature for 15 min or 100% methanol at −20 °C for 10 min, permeabilized in 0.2% Triton X-100, briefly washed with PBS three times, blocked with 1% BSA in PBS for 30 min, and probed with primary antibodies. Antibody–antigen complexes were detected with Alexa Fluor 594- or Alexa Fluor 488-conjugated goat secondary antibodies (ThermoFisher) by incubation for 30 min at room temperature. The cells were washed three times with PBS and then counterstained with 4′, 6′-diamidino-2-phenylindole (DAPI). Immunostained cells were examined under a fluorescence microscope (Zeiss Axioskop2; Carl Zeiss Microimaging Inc.) equipped with a Zeiss ApoTome system (Carl Zeiss Microimaging Inc.) and a confocal microscope (LSM800; Carl Zeiss Microimaging Inc.).

Spindle-orientation analysis

Z-stack images of metaphase cells stained with anti-pericentrin and anti-α-tubulin antibodies were scanned at 0.3-μm intervals using a confocal microscope. The xy-plane linear distance and the z-plane vertical distance between the two poles of the spindle in a series of the Z-stack images were measured using microscope software (ZEN; Carl Zeiss Microimaging Inc.). The spindle angle was calculated using the following formula: tan⁻¹(z distance/xy distance).

In vitro phosphorylation assay

Sixty nanograms of His-PLK1 (Sigma Aldrich) was incubated with 10 mM ATP and 2 μg of GST-WDR62 or GST-WDR62-S897A purified using glutathione-Sepharose 4B (GE Healthcare) following the manufacturer’s protocol in kinase buffer (25 mM MOPS pH 7.2, 25 mM MgCl₂, 5 mM EGTA, 2 mM EDTA, 50 mM DTT, 12.5 M β-glycerol phosphate, 5 mM ATP) for 2 h at 30 °C. The boiled samples were analysed by SDS-PAGE and western blotting.

Cytogenetics

Cells were treated with 0.1 μL/mL colcemid for 1 h. The cells were washed with Hanks balanced salt solution, trypsinized, and incubated in hypotonic solution (0.075 M KCl) for 20 min at room temperature. After centrifugation to collect the cells, they were fixed in Carnoy solution (3:1 ratio of methanol:acetic acid) for 1 h at room temperature. The fixed cells were then spread on slide glasses using HANABI Metaphase Spreader (Adstec Co.) and stained with 5% Giemsa solution (Sigma Aldrich) in Sorensen buffer (3:2 ratio of KH₂PO₄:Na₂HPO₄). Automatic scanning of metaphase spreads was performed by MetaSystems Scanning and Imaging Platform (Metapher4, Carl Zeiss AxioImager.Z2), and then the karyotype was analysed in 100 metaphases by semi-automatic counting of chromosomes in Ikaros Karyotyping Platform (MetaSystems).

Statistical analysis

The experiments were performed independently three times and the data are shown as mean ± s.d. Differences between groups were evaluated for statistical significance using Student’s t-test. Values of P < 0.05 were considered to be statistically significant.

Supplementary Material

Supplementary Material is available at HMG online.

Acknowledgements

We thank Dr Takashi Toda, Dr Hiroshi Ochiai, Dr Ken-ichi Suzuki and Dr Akihito Inoko for helpful discussions. We also thank Ms Yukiji Tonouchi and Ms Yuka Uchikawa for technical assistance.

Conflict of Interest statement. None declared.

Funding

Ministry of Education, Culture, Sports, Science and Technology of Japan (to S.M. and T.M.); the Center of World Intelligence Projects for Nuclear S&T and Human Resource Development from the Japan Science and Technology Agency (to S.M.); a Grant-in-Aid for Scientific Research from the Ministry of Health, Labor and Welfare (to S.M.); AMED-PRIME from the Japan Agency for Medical Research and Development, AMED (to T.M.); research grants from the Naito Foundation (to S.M. and T.M.); the Mochida Memorial Foundation for Medical and Pharmaceutical Research (to T.M.); Ono Medical Research Foundation (to T.M.); Takeda Science Foundation (to T.M.); a Phoenix Leader Education Program Research Grant (to S.A.).

Author notes

†

These authors contributed equally to this work.