Truncated SALL1 Impedes Primary Cilia Function in Townes-Brocks Syndrome

Abstract

Townes-Brocks syndrome (TBS) is characterized by a spectrum of malformations in the digits, ears, and kidneys. These anomalies overlap those seen in a growing number of ciliopathies, which are genetic syndromes linked to defects in the formation or function of the primary cilia. TBS is caused by mutations in the gene encoding the transcriptional repressor SALL1 and is associated with the presence of a truncated protein that localizes to the cytoplasm. Here, we provide evidence that SALL1 mutations might cause TBS by means beyond its transcriptional capacity. By using proximity proteomics, we show that truncated SALL1 interacts with factors related to cilia function, including the negative regulators of ciliogenesis CCP110 and CEP97. This most likely contributes to more frequent cilia formation in TBS-derived fibroblasts, as well as in a CRISPR/Cas9-generated model cell line and in TBS-modeled mouse embryonic fibroblasts, than in wild-type controls. Furthermore, TBS-like cells show changes in cilia length and disassembly rates in combination with aberrant SHH signaling transduction. These findings support the hypothesis that aberrations in primary cilia and SHH signaling are contributing factors in TBS phenotypes, representing a paradigm shift in understanding TBS etiology. These results open possibilities for the treatment of TBS.

Keywords: BioID; SALL1; Townes-Brocks syndrome; primary cilia; rare disease; spalt.

Figure 1

TBS^{p.Pro332Hisfs∗10} Cells Show Aberrant Cilia Frequency and Length (A and B) Micrographs of control HFFs (A) or TBS^{p.Pro332Hisfs∗10} cells (TBS³³²) (B) analyzed during cilia assembly and disassembly. Cilia were visualized by ARL13B (purple), basal bodies were visualized by pericentrin (green), and nuclei were visualized by DAPI (blue). (C and D) Graphical representation of cilia frequency (C) and cilia length (D) measured in control HFFs (blue circles; n = 58 cilia) or TBS³³² cells (orange triangles; n = 116 cilia) from three independent experiments. Cells that underwent 48 hr of starvation were compared with non-starved cells. After starvation, cells were supplied with serum for 30 min or 2 hr (n = 21–30 micrographs for all cases). (E) Micrographs of RPE1 cells infected with lentivirus expressing GFS alone (left) or GFS-SALL1^c.826C>T (GFS-SALL1²⁷⁶; right) and labeled with ARL13B (purple) and DAPI (blue). (F) Graphical representation of cilia length measured in control (blue circles; n = 43 cilia) or GFS-SALL1²⁷⁶ (orange triangles; n = 36 cilia) RPE1 cells infected as in (E). Graphs represent mean and SEM of three independent experiments. p values were calculated with a two-tailed unpaired Student’s t test: ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001. Scale bars, 5 μm.

Figure 2

Fibroblasts Derived from an Additional TBS Individual Exhibit Cilia Anomalies (A) Clinical findings in individual TBS^p.Arg276∗ (TBS²⁷⁶). (B) Western blot analysis of lysates from control ESCTRL#2 and TBS²⁷⁶. Samples were run in duplicate on the same gel and probed against SALL1^FL (asterisk, R&D antibody) or against SALL1²⁷⁶ (black arrowhead, N-terminal specific antibody⁶) TBS²⁷⁶ cells encoded with a ∼48 kDa truncated protein that positively reacts against SALL1 antibody. Vinculin and GAPDH were used as loading controls. Molecular-weight markers are shown to the right. (C and D) Confocal micrographs showing endogenous SALL1^FL localization in control ESCTRL#2 (C) or TBS²⁷⁶ (D) fibroblasts detected by FL-specific antibody (R&D). DAPI was used to label the nuclei (blue), and black and white images show the single green channel. Scale bars, 5 μm. (E and F) Immunofluorescence micrographs showing cilia marked with acetylated tubulin (purple), basal bodies marked with ODF2 (green), and nuclei marked with DAPI (blue) in control ESCTRL#2 and TBS²⁷⁶ fibroblasts. Pictures were taken with a Zeiss Axioimager D1 fluorescence microscope with a 63× objective. Scale bars, 2.5 μm. (G and H) Graphical representation of cilia length measurements (G; n = 123 and 235 cilia for control ESCTRL#2 and TBS²⁷⁶ fibroblasts, respectively) and cilia frequency (H; n = 28 and 31 micrographs for control ESCTRL#2 and TBS²⁷⁶ fibroblasts, respectively) of micrographs as shown in (E) and (F). Three independent experiments were pooled together. The graphs represent the mean and SEM. The p values were calculated with a two-tailed unpaired Student’s t test: ^∗∗∗p < 0.001. (I and J) Confocal micrographs showing SALL1^FL-YFP localization in control ESCTRL#2 and TBS²⁷⁶ fibroblasts. Actin is labeled by phalloidin (purple), and nuclei are labeled by DAPI (blue). Black and white images show the single green channel. Scale bars, 5 μm.

Figure 3

TBS^{p.Pro332Hisfs∗10} Fibroblasts Show Aberrant SHH Signaling (A and B) Graphical representation of the fold change of activation based on the relative expression of GLI1 (A) and PTCH1 (B) obtained by qPCR from control HFFs (n = 4; blue bars) or TBS^{p.Pro332Hisfs∗10} fibroblasts (TBS³³²) (n = 4; orange bars) either treated or not (−) with purmorphamine for the indicated times. Numerical values for the fold change are indicated. (C) Western blot analysis of lysates from control ESCTRL#2 and TBS^p.Arg276∗ (TBS²⁷⁶). Samples were probed against GLI3 (A, activator; R, repressor), and GAPDH was used as a loading control. Molecular-weight markers are shown to the right. (D) Graphical representation is shown for the fold change of the GLI3-A:GLI3-R ratios obtained in (C) for control and TBS fibroblasts. Data from at least three independent experiments using ESCTRL#2 (control; n = 3; blue bars) and TBS (TBS^276∗ and TBS³³²; pooled; n = 5; orange bars) are shown. (E) Graphical representation of the fold change in luciferase activation when Shh-LIGHT2 control cells (n = 8; blue bars) or cells expressing the mutated form GFS-SALL1^c.826C>T (SALL1²⁷⁶; n = 8; orange bars) were either treated (+) or not (−) with purmorphamine. All graphs represent the mean and SEM. p values were calculated with a two-tailed unpaired Student’s t test: ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001.

Figure 4

CCP110 and CEP97 Interact with Truncated SALL1 (A and B) Confocal micrographs showing transient expression of transfected BirA^∗-SALL1^FL (A) or BirA^∗-SALL1^c.826C>T (BirA^∗-SALL1²⁷⁶) (B) detected by fluorescence streptavidin. Nuclei are marked by DAPI in blue. Single green channels are shown in black and white (A′ and B′). Scale bars, 5 μm. (C) Venn diagram showing the distribution of the identified candidates by MS analysis. 665 and 1,032 proteins were found in close proximity of SALL1^FL and SALL1^p.Arg276∗ (SALL1²⁷⁶), respectively, in at least two independent experiments. (D) Volcano plot representing the distribution of the candidates identified by proximity proteomics in at least two out of three independent experiments. Proteins with at least a 2-fold change in intensity with respect to the WT proteome (log₂ ≥ 1) were considered SALL1²⁷⁶-associated candidates (orange dots). Proteins with less than a 2-fold change in intensity with respect to the WT proteome (log₂ ≤ 1) were considered SALL1^FL associated candidates (blue dots). (E) Graphical representation of the −log₁₀ of the p value for each of the represented cellular component GO terms. MT, microtubule; Extracell, extracellular. (F) Western blot of inputs or GFP-Trap pull-downs performed in HEK293FT cells transfected with SALL1^FL-YFP (lanes 1 and 6), SALL1^c.826C>T-YFP (SALl1²⁷⁶-YFP; lanes 2 and 7), SALL1²⁷⁶-YFP together with SALL1^FL-2xHA (SALL1^FL-HA; lanes 3 and 8), SALL1^FL-HA together with YFP alone (lanes 4 and 9), or YFP alone (lanes 5 and 10). Endogenous CCP110 and CEP97 were detected with specific antibodies. GAPDH was used as a loading control. Numbers under the CCP110 and CEP97 panels are the result of dividing the band intensities of each lane by that of lane 1 for the inputs or by that of lane 6 for the pull-downs. One asterisk indicates SALL1^FL-YFP, one black arrowhead indicates SALL1²⁷⁶-YFP, and two black arrowheads indicate YFP alone. Molecular-weight markers are shown to the right.

Figure 5

TBS^{p.Pro332Hisfs∗10} Cells Show Changes in the Localization of CCP110 and CEP97 (A and C) Immunofluorescence micrographs of cycling human-derived fibroblasts stained with antibodies against endogenous CCP110 (A) or CEP97 (C) (green), CEP164 to label the mother centriole (MC; purple), and DAPI to label the nuclei (blue). Black and white images show the single green and purple channels. Note the different distribution of CCP110 and CEP97 to the MC between TBS^{p.Pro332Hisfs∗10} fibroblasts (TBS³³²) and control HFFs. (B and D) Graphical representation of the percentage of cells showing the presence of CCP110 or CEP97 at the MC per micrograph corresponding to the experiments in (A) or (C), respectively; n = 30 micrographs. Three independent experiments were pooled together. Pictures were taken with a Zeiss Axioimager D1 fluorescence microscope with a 63× objective. All graphs represent the mean and SEM. Scale bars, 1 μm.

Figure 6

Truncated SALL1 Leads to a TBS-like Cilia Phenotype in Genetically Modified Cells (A) Western blot analysis of lysates from WT HEK293FT cells (293^WT) or CRISPR/Cas9 cells (293³³⁵ cells). The latter were modified to mimic a TBS mutation in SALL1 (SALL1^c.1003dup). A truncated protein of about 48 kDa was identified in 293³³⁵ cells by SALL1 antibody (SALL1^{p.Ser335Lysfs∗20}; arrowhead). The asterisk indicates SALL1^FL. GAPDH was used as a loading control. Molecular-weight markers are shown to the right. (B) Micrographs of immunofluorescence using a specific antibody (green) show the subcellular distribution of SALL1^FL in control or mutated cells. ARL13B marks the primary cilium (purple and yellow arrowheads), and DAPI marks the nuclei. Yellow circles indicate the nuclei, and black and white images show the single green channel. Scale bars, 5 μm. (C and D) Graphical representation of cilia frequency (C) and cilia length measurements (D) of micrographs shown in (B) (cilia frequency: n = 11 micrographs; cilia length: n = 38 cilia in 293^WT cells and n = 48 cilia in 293³³⁵ cells). Three independent experiments were pooled together. The graphs represent the mean and SEM. The p values were calculated with a two-tailed unpaired Student’s t test: ^∗∗p < 0.01, ^∗∗∗p < 0.001. (E and F) Immunofluorescence micrographs of 293^WT or 293³³⁵ cells showing centrosomes (purple), CCP110 foci (green), and nuclei (DAPI in blue). Scale bars, 1 μm. Pictures were taken with a Zeiss Axioimager D1 fluorescence microscope with a 63× objective. (G) Graphical representation of the percentage of cells showing CCP110 in two foci per micrograph in 293^WT or 293³³⁵ cells, corresponding to the experiments in (E) or (F), respectively.

Figure 7

Cells Derived from Sall1^+/Δ Mouse Embryos Exhibit Cilia Defects (A) PCR genotyping of MEFs derived from WT homozygous (Sall1^+/+) or Sall1-ΔZn^2–10 heterozygous (Sall1^+/Δ) embryos. (B) Micrographs of Sall1^+/+ and Sall1^+/Δ MEFs analyzed during cilia assembly. Cilia were visualized by acetylated alpha-tubulin and gamma-tubulin (AcTub and γTub, respectively; purple), and nuclei were visualized by DAPI. Scale bar, 1 μm. (C and D) Graphical representation of cilia length (C) and cilia frequency (D) measured in Sall1^+/+ (n = 38 cilia and 24 micrographs for cilia length and frequency, respectively; blue circles) and Sall1^+/Δ MEFs in (B) (n = 37–45 cilia and n = 16–38 micrographs for cilia length and frequency, respectively; orange triangles). (E) Immunofluorescence micrographs of Sall1^+/+ and Sall1^+/Δ MEFs showing the centrosomes (AcTub, purple), CCP110 foci (green), and nuclei (DAPI). Scale bars, 0.5 μm. Cells that underwent 48 hr of starvation were compared with non-starved cells. Pictures were taken with a Deltavision fluorescence microscope (GE Healthcare Life Sciences) with a 60× objective. (F) Graphical representation of the percentage of cells showing CCP110 in two foci in Sall1^+/+ and Sall1^+/Δ MEFs in (E) (n > 10). The graphs represent the mean and SEM. p values were calculated by one-way ANOVA: ^∗∗p < 0.01, ^∗∗∗p < 0.001. (G) The presence of a truncated SALL1 underlies cilia malformations in TBS. In control cells (left), SALL1^FL is mainly nuclear, and CCP110 and CEP97 localize in the mother and daughter centrioles, inhibiting cilia formation. Upon starvation, CCP110 and CEP97 are depleted from the MC, which will allow the formation of the primary cilia. By contrast, in TBS cells (right) a truncated form of SALL1, together with sequestered SALL1^FL, localizes in the cytoplasm, where it interacts with centrosomal proteins such as CCP110 and CEP97, displacing them from the MC. As a result, the frequency of cilia formation increases, and cilia are longer than in control cells. Problems in cilia assembly are accompanied by altered SHH signaling in TBS cells.