Fig 2: Identification of MIF target genes with ChIP-seq analysis.a The distribution of MIF ChIP-peaks was across different genomic regions. The pie chart shows that MIF binds to the promoters-TSS (transcription start site) region in about 14% of its target genes. b The distribution of promoter peaks identified with MIF ChIP-seq was markedly concentrated near the TSS. c, d KEGG pathway and GO categories analysis of the MIF target genes. Peaks of MIF occupancy on the promoters of KIF3a (e), IFT20 (f), CEP290 (g) and BBS4 (h) were identified by manual inspection and were shown as the input and MIF ChIP.

Fig 3: Dual-color STED images revealing distinct axial localization levels of different TZ and TF proteins.(a–f) Sample dual-color STED images showing different axial positions relative to the distal edge of centrin for different TZ/TF proteins. The axial distance of each BD V500-labeled TZ/TF protein to the FWHM-defined boundary of centrin-eGFP was measured. Centrin-eGFP signal could also be seen in the BD V500 channel. (g–l) Comparison of the axial positions of TZ/TF proteins (green) relative to centrin (yellow) revealing that TMEM67 and TCTN2 were dwelled at a similar axial level to MKS1 and RPGRIP1L in terms of the distance to centrin (red dashed line), while CEP290 was at another axial level close to centrin, distinct from the other TZ proteins. CEP164 was slightly proximal to the distal edge of centrin. Scale bar for (a–l): 200 nm. (m) Comparison of the axial distances to centrin. RPGRIP1L, MKS1, TMEM67, and TCTN2 were ~150 nm from the centrin edge, while CEP290 was ~40 nm from the centrin edge. (n) A series of single-color STED images of TZ/TF proteins were axially positioned based on the average relative distances to the distal edge of centrin from dual-color STED images. The threshold of each single-color image was adjusted based on the FWHM to represent their shape and size. These axially-positioned images were overlapped to create a 7-color superresolution image of proteins at the ciliary base. To obtain a white-background image to be merged with an EM image, colors of the merged multi-color image were inverted with a corresponding negative mapping (upper). Scale bar: 200 nm.

Fig 4: Cep290 does not require Ahi1 for localisation to centrosomes.Immunofluorescence of IMCD3 monolayers co-transfected with siGLO and either negative-control siRNA (A and B), or siRNA against Ahi1 (A and C). (A) Graph to show centrosome position (gamma tubulin immunofluorescence, graph expressed as mean +/− s.e.m.) from apical to basal in control and Ahi1-silenced cells. Apical refers to the upper 1.5–2 µm of the cell (the distance between the nuclear envelope and the plasma membrane in these cells; this is unchanged following silencing of Ahi1, data not shown). Arrows denote the apical and mid-cell positions used in B and C. (B and C) Confocal maximum intensity projections of control (B) and Ahi1-silenced (C) cells showing Cep290 (red), gamma tubulin (green) and DNA (blue). Transfected cells were identified using siGLO (yellow). In each case, the top panel represents an apically-oriented centrosome (position 1 on the graph, A) and the bottom panel represents a centrosome from the mid-cell region (position 2 on the graph). Scale bar: 5 µm.

Fig 5: Variable penetrance of the CDT phenotype in tmem67, rpgrip1l and cep290 mutants. (A) Lateral views of WT and mutants at 2 dpf and 5 dpf. All homozygous mutant embryos were obtained from a single pair of heterozygous mutant crosses that were dechorionated at 1 dpf. Embryos were sorted for CDT at 2 dpf and then reanalyzed at 5 dpf. All embryos were genotyped after phenotype analysis at 5 dpf. Arrows denote the phenotypic changes across time. Percentage was calculated as the % of observed number of −/− at 5 dpf divided by the observed number of −/− at 2 dpf. Scale bars: 1000 μm. (B) Bar graph representing distribution of CDT phenotype at 2 dpf and 5 dpf. For tmem67−/−, n=274 from ten breeding pairs of tmem67+/− adults. For cep290−/−, n=351 from 23 breeding pairs of cep290+/− adults. For rpgrip1l−/−, n=226 from ten breeding pairs of rpgrip1l+/− adults. Arrows point to derivates of the 2 dpf phenotype. Error bars represent ±s.e.m. (C) Table depicting the revertant rate from repeatedly bred cep290+/− pairs. (D) Table depicting the ratio of embryos with normal-looking tail from repeatedly bred rpgrip1l−/− pairs. As above, embryos were dechorionated at 1 dpf, sorted at 2 dpf, reanalyzed at 5 dpf and genotyped after. Data are only for cep290−/− and rpgrip1l−/− embryos. Red boxes in C and D denote pairs that had repeatedly high reversion rates. (E) Comparison of spinal canal cilia in tails of cep290l−/− embryos with a CDT or reverted normal-looking tail at 5 dpf. Representative brightfield images of the spinal canal and fluorescent images of acetylated-Tubulin staining (green fire blue LUT) are shown for WT, cep290−/− embryos with a normal-looking tail and cep290−/− embryos with CDT. Dashed lines outline approximate boundaries of the spinal canal. Scale bars: 25 μm. (F) Comparison of spinal canal cilia in tails of rpgrip1l−/− embryos with CDT or normal-looking tail. Representative brightfield images of the spinal canal and fluorescent images of acetylated-Tubulin staining (green fire blue LUT) are shown for rpgrip1l+/+ and rpgrip1l−/− embryos with a normal-looking (straight) tail, and rpgrip1l−/− embryos with a CDT. Dashed lines outline approximate boundaries of the spinal canal. n=number of embryos with the depicted staining/total number of embryos analyzed. Scale bars: 25 μm. (G) K-means clustering of log2 normalized counts define modifier signature cluster among the top 100 DEGs in 2 dpf rpgrip1l−/− with CDT, rpgrip1l−/− with normal-looking tail (normal) and rpgrip1l+/+ groups (WT).