The kinetochore protein, CENPF, is mutated in human ciliopathy and microcephaly phenotypes

Abstract

Background: Mutations in microtubule-regulating genes are associated with disorders of neuronal migration and microcephaly. Regulation of centriole length has been shown to underlie the pathogenesis of certain ciliopathy phenotypes. Using a next-generation sequencing approach, we identified mutations in a novel centriolar disease gene in a kindred with an embryonic lethal ciliopathy phenotype and in a patient with primary microcephaly.

Methods and results: Whole exome sequencing data from a non-consanguineous Caucasian kindred exhibiting mid-gestation lethality and ciliopathic malformations revealed two novel non-synonymous variants in CENPF, a microtubule-regulating gene. All four affected fetuses showed segregation for two mutated alleles [IVS5-2A>C, predicted to abolish the consensus splice-acceptor site from exon 6; c.1744G>T, p.E582X]. In a second unrelated patient exhibiting microcephaly, we identified two CENPF mutations [c.1744G>T, p.E582X; c.8692 C>T, p.R2898X] by whole exome sequencing. We found that CENP-F colocalised with Ninein at the subdistal appendages of the mother centriole in mouse inner medullary collecting duct cells. Intraflagellar transport protein-88 (IFT-88) colocalised with CENP-F along the ciliary axonemes of renal epithelial cells in age-matched control human fetuses but did not in truncated cilia of mutant CENPF kidneys. Pairwise co-immunoprecipitation assays of mitotic and serum-starved HEKT293 cells confirmed that IFT88 precipitates with endogenous CENP-F.

Conclusions: Our data identify CENPF as a new centriolar disease gene implicated in severe human ciliopathy and microcephaly related phenotypes. CENP-F has a novel putative function in ciliogenesis and cortical neurogenesis.

Keywords: CENPF; Ciliopathy; Clinical genetics; Microcephaly; Molecular genetics.

Figure 1

Pedigrees and clinical features of families with CENPF mutations. (A) The pedigree shows Family 1 with novel ciliopathy disorder consisting of non-consanguineous unaffected parents with six offspring, of which four were affected and died in utero, and two were unaffected and are healthy. (B) Gross morphological features of an affected fetus with dysmorphic craniofacial features such as a high nasal bridge, short columella, micrognathia, wide mouth and low-set ears. Examination of a dissected gastrointestinal tract from the same affected fetus revealed complete duodenal atresia. (C) The pedigree of Family 2 with a single affected case of microcephaly (MCPH) and unaffected parents and two unaffected siblings. (D) Chart demonstrating the occipital head circumference (OFC) data for the affected fetuses of kindred 1 (1.1, 1.3 and 1.4, at or below 3rd percentile) at gestational age at the time of autopsy in addition to the OFC at birth for the patient with MCPH of Family 2 which is below the 3rd percentile at 40 weeks gestation.

Figure 2

(A and B) CENPF genomic organisation, depicting locations of the identified heterozygous essential splice site non-synonymous mutation, IVS5-2A>C, and the heterozygous non-synonymous non-sense mutations, c.1744G>T and c.8692C>T (blue). For details on segregation, see also online supplementary figures S3 and S4. (C) CENPF encodes a protein of 350 kDa, consisting of 3114 amino acid residues. CENP-F protein consists of mainly coiled coil domains (blue), several leucine heptad repeats (purple), microtubule (MT)-binding domains at both the N and C termini in addition to Nudel (Nde) binding, kinetochore (KT)-binding and Nup133-binding domains. The kinetochore localisation domain and a bipartite nuclear localisation sequence reside in the C-terminal region.

Figure 3

CENP-F is expressed at basal bodies of ciliated cells. (A) Shown are representative micrographs of cilia following dual immunofluorescence labelling of ciliated NIH 3T3 fibroblasts with anti-CENPF and anti-α-acetylated tubulin antibodies which demarcates cilia. CENP-F is localised to the basal bodies (arrows) of ciliated NIH 3T3 fibroblasts. Scale bar, 10 μm. (B) CENP-F colocalises with Ninein (arrows) at the subdistal appendages of the mother centriole of ciliated IMCD3 cells. Scale bar 5 μm. (C) Ultrastructural localisation of CENP-F in serum-starved retinal pigmentary epithelial cells. Black arrows point to immunogold particles along the microtubules and subdistal appendages of the mother centriole. Scale bar 100 nm.

Figure 4R2

(A) Zebrafish cenpf morphants display increased body axis curvature at 24 h postfertilisation (hpf) compared with control embryos (black arrow). Cenpf knockdown in cardiac myosin light chain (cmlc2)-gfp transgenic zebrafish causes laterality heart defects at 48 hpf. Hydrocephalus (arrow) is evident at 72 hpf in cenpf morphants compared with control embryos. At 96 hpf, pronephric cysts (arrow) are evident in cenpf morphants compared with control embryos. (B) Quantitative graph showing increased occurrence of axis curvature defects, laterality malformations, hydrocephalus and pronephric cysts in cenpf morphants (blue bars) compared with control embryos (red bars) and compared with cenpf morphants injected with human CENPF RNA (black bars). Bars represent an average of three experiments. Error bars denote SE of the mean (SEM). [Std-MO (n=266) % ventral axis curvature at 24 hpf vs cenpf-MO (n=173) 12.7±1.5 vs 88.7±1.4, *p<0.001; cenpf-MO (n=173) vs cenpf-MO with human CENPF RNA (n=256) 88.7±1.4 vs 38.7±2.0, *p<0.001; Std-MO (n=223) % laterality defects at 48 hpf vs cenpf-MO (n=152) 4.0±0.6 vs 81.7±2.8, *p<0.001; cenpf-MO (n=152) vs cenpf-MO with human CENPF RNA (n=229) 81.7±2.8 vs 28±2.6, ***p<0.01; Std-MO (n=204) % hydrocephalus at 72 hpf vs cenpf-MO (n=93) 1±0.6 vs 68.3±1.7, *p<0.001; cenpf-MO (n=93) vs cenpf-MO with human CENPF RNA (n=197) 68.3±1.7 vs 40.3±1.7, *p<0.001; Std-MO (n=158) % pronephric cysts at 120 hpf vs cenpf-MO (n=76) 1.2±0.9 vs 96±0.6, ****p<0.0001; cenpf-MO (n=76) vs cenpf-MO with human CENPF RNA (n=122) 96±0.6 vs 37.3±1.9, *p<0.001]. (C) Representative images of southpaw mRNA expression in the lateral plate mesoderm at 18-somites (ss) of control (a) and cenpf morphant embryos (b-d). (a) left-sided expression in control embryos (arrow, top left panel). (b) right-sided expression (arrow), (c) bilateral expression and (d) absent expression in stage-matched cenpf morphant embryos (arrows). Scale bar 50 μm. (D) Representative micrographs following immunofluorescent labelling of Kupffer's vesicle (KV) cilia with anti-α-acetylated tubulin antibody at 8 ss. Short KV cilia are noted in cenpf morphants (white arrows) (E) Quantitative graph showing a quantitative difference in KV cilia length (µm) in cenpf morphants (n=136 cilia; n=5 embryos) vs controls (SD MO) (n=228 cilia; n=4 embryos); 4.2±0.4 vs 2.6±0.1 **p<0.0001). (F) Quantitative graph showing that KV cilia number were significantly less in cenpf morphants (n=5 embryos) vs controls (std MO) (n=5 embryos); 56.4±1.9 vs 38.6±1.7 **p<0.001). (G) Long cilia are observed in the lumina of collecting ducts of control fetuses (white arrow) while short cilia are evident on renal epithelial cells of CENPF mutant fetal kidneys (white arrow). Sections are counterstained with 4′,6-diamidino-2-phenylindole. Scale bar 10 μm. MO, morpholino oligonucleotides.

Figure 5

(A) Colocalisation of CENP-F at the centrosome with intraflagellar transport-88 (IFT88) (B) Representative micrographs of asynchronous 3T3 fibroblasts following dual immunofluorescent labelling of 3T3 cells with KIF3B and C-terminus CENP-F antibody. CENP-F localises to the centrioles with KIF3B (arrows). Scale bar 5 μm. Inset: high-power view of CENP-F localisation between two KIF3B foci. (C) Co-localisation of CENP-F along ciliary axonemes labelled with IFT88 antibody (D) IFT88 localises to long cilia within the lumina of renal collecting ducts of 22-week-old control human fetuses (arrows). Inset shows colocalisation of IFT88 with GT335-positive ciliary axonemes in xy and xz plane of confocal projection. Scale bar xy, 15 µm. (E) IFT88 does not localise with GT335-positive ciliary axonemes, if present, of CENPF mutant fetal kidneys (arrows). GT335-positive cilia are shorter than cilia in control kidneys. Scale bar 10 μm. Inset shows that IFT88 and GT335 do not colocalise in the xy and xz plane of confocal projection. Scale bar 10 μm. (F) Representative images of co-immunoprecipitation experiments carried out on protein lysates from mitotic HEKT293 cells containing endogenous CENP-F. Immunoblots show that IFT88, KIF3B and CENP-F co-immunoprecipitate with endogenous CENP-F, while an immunoglobulin G (IgG) isotype control does not co-immunoprecipitate with CENP-F. IN=input; 10% of total input is indicated. (b) Reciprocal co-immunoprecipitation experiments carried out on protein lysates from serum-starved HEKT293 cells containing endogenous IFT88. Immunoblots show that CENP-F co-immunoprecipitates with endogenous IFT88, while an IgG isotype control does not co-immunoprecipitate with IFT88. IN=input; 10% of total input is indicated. (G) Asynchronous HeLa cell lysate was fractionated over a superose-6 gel filtration column. Eluted fractions were probed with antibodies against CENP-F, IFT complex B members: IFT88, IFT52 and IFT20, and motors: cytoplasmic dynein 1 intermediate chain (Dyn IC 74.1) and Kif3a. CENP-F co-eluted with the IFT proteins and motors, suggesting that it exists as a complex with these proteins. Arrows indicate peak elution fractions for calibration proteins: thyroglobulin (669 KDa; fraction 23), β-amylase (200 KDa, fraction 28) and bovine serum albumin (BSA) (67 KDa, fraction 31). V, void volume.