Wdr60 and Wdr34 genes were disrupted by piggyBac (PB) transposon system in mice respectively. As shown in Figure 1A, a DNA fragment containing RFP sequence was inserted between exon 4 and 5 of Wdr60 gene. We firstly tested the efficiency of PB insertion in Wdr60 ^PB/PB mice. There are four predicted transcriptional variants for Wdr60, which are indicated as transcript X1, X2, NM, and X5 in Figure 1A. The transcript NM (NM_146,039.3) is the functional Wdr60 transcript, and it was significantly downregulated by PB insertion in Wdr60 ^PB/PB homozygotes without affecting the neighboring genes Esyt2 and Vpr2 (Figure 1B). The protein encoded by this functional transcript NM is around 100 kDa in mice. To confirm the decrease of WDR60 protein, we extracted total protein from E14.5 heads and immunoblotted with anti-WDR60 antibody. As shown in Figures 1C, D, the band at 100 kDa was significantly decreased in Wdr60 ^PB/PB mice compared to that in heterozygote (HET) mice. We also noticed a decrease of the 75 kDa band in the blot (Figure 1C). Therefore, we performed mass spectrometry on both 100 and 75 kDa bands. WDR60 is one of the proteins detected by mass spectrometry at 100kDa, while none of the 75 kDa proteins detected by mass spectrometry is WDR60 (Supplementary Table S1). Although other alternative transcripts of Wdr60 was upregulated in Wdr60 ^PB/PB mice (Figure 1B), they are probably not translated as no other protein isoform was detected. Immunofluorescent experiment was performed on E10.5 wild type (WT) embryo with anti-WDR60 antibody, and the results revealed that WDR60 is mostly expressed at the brain and heart regions at E10.5 (Figure 1E). Meanwhile, no fluorescent signal can be detected in Wdr60 ^PB/PB mice with anti-WDR60 antibody (Figure 1F). Therefore, the expression of WDR60 was efficiently disrupted by PB insertion in Wdr60 ^PB/PB mice.

The gene structure of Wdr34 is relatively simpler and has only one transcript. Wdr34 gene was interrupted by PB insertion between exon 1 and 2 (Figure 1G). The expression of Wdr34 transcript is also significantly downregulated in Wdr34 ^PB/PB mice without affecting the neighboring genes (Figure 1H). Hence, both Wdr60 and Wdr34 are knocked down efficiently in our models.

Both Wdr60 ^+/PB and Wdr34 ^+/PB heterozygote mice appeared normal. Both Wdr60 ^PB/PB and Wdr34 ^PB/PB mice are embryonic lethality with a complete penetrance based on the results from heterozygotes breeding. Genotyping results of the offspring from heterozygotes breeding indicated that Wdr60 ^PB/PB and Wdr34 ^PB/PB embryos die at different stages (Tables 1, 2). Wdr60 ^PB/PB mice die between E13.5 to E14.5 (Table 1), while Wdr34 ^PB/PB mice die during E10.5 to E11.5 (Table 2 ). Most Wdr60 ^PB/PB embryos exhibit abnormal head development at E10.5 with smaller telencephalic evagination and midbrain (Figure 2A, indicated by read arrow head and white star). The maxillary processes in Wdr60 ^PB/PB embryos are more visible compared to that in WT embryos at E10.5 (Figure 2A). Some of the Wdr60 ^PB/PB embryos also have cardiac edema and twisted tail at E10.5 (Figure 2A; Table 3). On the other hand, Wdr34 ^PB/PB mice present severe developmental retardation and malformation at E10.5 (Figure 2B; Table 4). Similar to Wdr60 ^PB/PB embryos at E10.5, some Wdr34 ^PB/PB embryos at E10.5 also have smaller telencephalic evagination and midbrain (Figure 2B, indicated by read arrow head and white star). Consistent with the previous reported Wdr34 null mice (Wu et al., 2017), some Wdr34 ^PB/PB embryos at E10.5 also have open head (Figure 2B, white triangle). However, there are more Wdr34 ^PB/PB embryos at E10.5 have severe growth retardation with an expanded heart loop and open caudal neural tube ( Figure 2B; Table 4).

We were able to collect some Wdr60 ^PB/PB embryos at E12.5 to E14.5. The ratio of total homozygotes at E12.5 still fit the Mendelian law. However, some of the homozygotes were dead at this stage. Therefore, the ratio of live homozygotes is less than expected at this stage (Table 1). The live Wdr60 ^PB/PB embryos at E12.5 displayed multiple defects, including smaller lower jaw, polydactyl, and edema (Figure 2C). Transverse sections at the head of the E12.5 embryos revealed that Wdr60 ^PB/PB embryos also have holoprosencephaly (Figure 2C).

We also performed immunofluorescent experiments on the transverse sections of E10.5 embryos with antibodies against acetyl-Tubulin, which is a cilia marker protein. High signal for acetyl-Tubulin can be observed in the neural tube, and cilia can be observed on the apical surface of the neural tube in WT embryos (Figures 2D, E). However, this signal was much less in the Wdr60 ^PB/PB embryos (Figure 2D) and it was not detectable in the Wdr34 ^PB/PB embryos (Figure 2E).

As WDR60 is highly expressed in the head at E10.5, we collected the head tissues of Wdr60 ^PB/PB and WT embryos at E10.5 for RNAseq experiment (Supplementary Table S2). Differential expressed genes are filtered and enriched by KEGG pathway analysis and Gene Ontology (GO) analysis. The results showed that both Hedgehog signaling pathway and ventral neural tube patterning were significantly downregulated in Wdr60 ^PB/PB mice (Figures 3A, B). The downregulated genes included Shh, Gli1, Hhip, Patched1, Patched2, Foxa1, Foxa2, Nkx6.1, Nkx6.2 and Nkx2.2 (Supplementary Table S2). We further confirmed the downregulation of Shh, Gli1, Hhip, Patched1, Patched2, Nkx6.1, and Nkx6.2 in Wdr60 ^PB/PB embryos by qRT-PCR experiments (Figure 3C). We also collected whole embryos from Wdr34 ^PB/PB and WT littermates at E10.5 for RNAseq experiment (Supplementary Table S3). SHH signaling was also downregulated in Wdr34 ^PB/PB embryos and the downregulation of Shh, Gli1, Patched1, Patched2, Nkx6.1, and Nkx6.2 was also confirmed by qRT-PCR in Wdr34 ^PB/PB embryos (Figure 3D). Therefore, both WDR60 and WDR34 are positive regulators in SHH signaling. SHH signaling is important for ventral neural tube patterning and it has been reported that Wdr34 ^−/− mice have dorsalized neural tube (Wu et al., 2017). We next examined the D-V neural tube patterning in Wdr60 ^PB/PB embryos by immunofluorescent experiment. The results showed that the neural tube patterning is also abnormal in Wdr60 ^PB/PB mice (Figure 4). The most ventral domains were not differentiated as neither NKX2.2 nor FOXA2 were detectable in Wdr60 ^PB/PB mice at two different transverse plains. The expression of NKX6.1 was still detectable in the Wdr60 ^PB/PB neural tube, but the positive region extended towards the floor plate in Wdr60 ^PB/PB mice at the heart level (Figures 4A–C, E). PAX6 positive region at this plane also extended to the floor plate in Wdr60 ^PB/PB mice (Figures 4A, D).

Both WDR60 and WDR34 are positive regulator in SHH signaling and both Wdr60 ^PB/PB and Wdr34 ^PB/PB embryos have ventral patterning defects in neural tube. However, this cannot explain the different penetrance of NTD phenotypes between Wdr60 ^PB/PB and Wdr34 ^PB/PB embryos as we observed much high ratio of open cranial and caudal neural tube in Wdr34 ^PB/PB embryos than that in Wdr60 ^PB/PB embryos (Tables 3, 4). The open neural tube phenotype is unlikely due to the defects in SHH signaling as it was downregulated in both Wdr60 ^PB/PB and Wdr34 ^PB/PB embryos. Previously, we have reported that WDR34 can regulated PCP signaling in vitro (Yin et al., 2020). PCP signaling has been reported to play essential role in neural tube closure through enhancing midline apical constriction and the convergent extension of neural tube (Nikolopoulou et al., 2017; Wang et al., 2019). Therefore, we examined the expression of c-Jun and Juk, two downstream targets of PCP signaling, in both Wdr34 ^PB/PB and Wdr60 ^PB/PB embryos by qRT-PCR. Our results showed that the expression of c-Jun is significantly downregulated in Wdr34 ^PB/PB embryos but not in Wdr60 ^PB/PB embryos (Figures 5A, B). In addition, we also examined several key components of PCP pathway in both Wdr34 ^PB/PB and Wdr60 ^PB/PB embryos. While the expression of Vangl1 and Vangl2 was downregulated in both homozygotes compared to WT embryos, the expression of Celsr1 and Celsr2 was only significantly downregulated in Wdr34 ^PB/PB embryos but not in Wdr60 ^PB/PB embryos (Figures 5A, B).

Further immunofluorescent experiments were performed on coronal sections of E10.5 embryos with antibodies against VANGL2 and CELSR1. The expression of VANGL2 was decreased in both Wdr60 ^PB/PB and Wdr34 ^PB/PB embryos compared to their WT littermates (Figure 5E). The expression of CELSR1 was not affected in Wdr60 ^PB/PB embryos; but it were significantly less in Wdr34 ^PB/PB embryos compared to their WT littermates (Figure 5F). These results correlate well with the results in qRT-PCR experiments.

We also examined the downstream targets of canonical WNT signaling in both Wdr34 ^PB/PB and Wdr60 ^PB/PB embryos by qRT-PCR. There was a small decrease on Axin2 in Wdr60 ^PB/PB embryos and on Lef1 in Wdr34 ^PB/PB embryos compared to their WT littermates (Figures 5C, D).

Previous study on cells have revealed that WDR60 and WDR34 have some distinctive functions in cilia as WDR60 but not WDR34 is responsible for dynein-2 assembly, whereas WDR34 but not WDR60 is essential for axonemal extension (Vuolo et al., 2018). We noticed that in that study, knockout of WDR60 or WDR34 seem to have different effect on IFT140 disruption in cilia as IFT140 was stuck at the cilia tip in Wdr34 knockout cells, but still remained at ciliary median shaft in Wdr60 knockout cells (Vuolo et al., 2018). It has been reported that Ift140 mutant mice also exhibit open brain phenotypes (Miller et al., 2013). Therefore, we hypothesized that IFT140 may be one of the key components different between WDR60 and WDR34. We first performed co-IP experiments to test whether WDR60 and WDR34 have different affinity with IFT140.293T cells were transfected with flag-tagged human WDR60 or WDR34 plasmids. IFT140 or IFT88 were pulled down by their specific antibody, and WDR60 or WDR34 pulled down together with IFT proteins were detected by western blot. Our co-IP results revealed that indeed only WDR34 but not WDR60 interact with IFT140 (Figures 6A, C). At the same time, both WDR60 and WDR34 interact with IFT88 (Figures 6A, B), which correlates with the similar distribution defects of IFT88 in Wdr60 and Wdr34 knockout cells (Vuolo et al., 2018). Since Ift140 mutant mice also exhibit open brain phenotypes (Miller et al., 2013), it is possible that IFT140 is involved in PCP signaling as well. Therefore, we tested whether IFT140 can regulate PCP signaling in human ARPE-19 and 293T cells. SiRNAs for IFT140 was transfected into 293T or ARPE-19cells and the expression of PCP downstream targets c-JUN, INTU, ATF2 and PCP pathway components VANGL1 and CELSR1 were detected by qRT-PCR. Our results showed that knockdown of IFT140 could downregulate PCP signaling in 293T cells as the expression levels of c-JUN and ATF2 were significantly downregulated in IFT140 knockdown cells (Figure 6D). However, no effect could be observed in ARPE-19 cells (Figure 6E).

The mice used for experiments were housed at the Animal center of the Institute of Developmental Biology and Molecular Medicine, Fudan University. FVB-Wdr60 ^PB/+ and C57BL-Wdr34 ^PB/+ strains were generated based on previously reported method (Ding et al., 2005). In brief, mice carrying a single PB transposon PB [Act-RFP] in the genome was mated with the transposase line Act-PBase (Ding et al., 2005), which were generated by conventional pronuclei injection of linear plasmids. Double-positive male progeny were then crossed with WT mice to generate transposase negative mutants carrying remobilized PB insertions. The Wdr60 and Wdr34 mutations were one of the mutagenesis lines generated in the offspring. The mutation was mapped by inverse polymerase chain reaction (PCR) and with the primers listed below. Wdr60 mutant carries the PB insertion in the third intron of the Wdr60 gene (Chr12:116,218,415 bp, Ensemble 109). Wdr34 mutant carries the PB insertion in the first intron of the Wdr34 gene (Chr2:29932629 bp, Ensemble 109).

The primers for genotyping Wdr60 are:

PB: 5′-CTG​AGA​TGT​CCT​AAA​TGC​ACA​GCG-3′,

GL: 5′-TCA​GAG​GTA​GTC​TTT​GCC​CAC​C-3′, and

GR: 5′-CCC​AAG​CTG​CTT​GTT​AGT​TTG​C-3’.

The primers for genotyping Wdr34 are:

PB: 5′-CTG​AGA​TGT​CCT​AAA​TGC​ACA-3′,

GL: 5′-GCT​ACA​TTT​CTG​GTT​AGT​CTG​GGT​C-3′, and

GR: 5′-TGT​CTA​GCA​CGG​CAC​ATG​CA-3’.

Images of mouse embryos were captured on Leica stereoscopic microscope. All animal experiments were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of Fudan University. The protocol was approved by the Committee on the Ethics of Animal Experiments of Fudan University.

Human WDR60 and WDR34 cDNA were cloned into pCMS-EGFP vector and the plasmids were generously provided by Prof. Zhiheng Xu of Chinese Academy of Science. The primers used for cloning WDR60 were 60-F-XhoI: 5′-CCG​CTC​GAG​ATG​GAG​CCC​GGG​AAG​A-3′ and 60-R-KpnI: 5′-AAT​TCT​AGA​TCA​GGC​CGC​CAC​CTC​TGC-3’. The primers used for cloning WDR34 were 34-F-EcoRI: 5′-ATT​AGA​ATT​CCA​TGG​CAA​CCC​GCG​CG-3′ and 34-R-XbaI: 5′-AAT​TCT​AGA​TCA​GGC​CGC​CAC​CTC​TGC-3’.

HEK-293T cells were grown in high-glucose DMEM or DMEM/F12 supplemented with 10% FBS and 1% penicillin/streptomycin in a 5% CO₂ incubator at 37°C. Human adult retinal pigmented epithelium-19 (ARPE-19) cells were grown in DMEM/F12 supplemented with 10% FBS and 1% penicillin/streptomycin in a 5% CO₂ incubator at 37°C. Transfection was performed using FuGENE^® HD Transfection Reagent (Promega) according to the manufacture’s instruction.

The head tissue of E10.5 mice were dissected and collected on ice. Total RNA of samples was extracted and fragmented for building cDNA libraries. The libraries were quality controlled and were sequenced on the Illumina HiSeq^® 2,500 System. The differential expressive genes were sorted by fold change >2 plus q-value <0.05, and were analyzed by KEGG pathway enrichment and Gene Ontology (GO) enrichment.

Embryos at stages were dissected on ice and oolemma of each individual was kept for genotyping. Total RNA was extracted with TRIzol reagent (Invitrogen) and chloroform, precipitated with isopropanol, washed by 70% RNase-free ethanol and reverse transcribed into cDNA. qRT-PCR experiments were performed using SYBR^® Green Realtime PCR Master Mix (Code No. QPK-201) on the LightCycler^® 480 Real-Time PCR System (Roche). Relative expression of each target was calculated with mouse Gapdh as control (Primers listed in Supplementary Table S4).

293T and ARPE-19 cells transfected with siRNAs were also lysed and total RNA was extracted as described above. Human GAPDH was used as internal control.

The sequences for siRNAs for IFT140 are as follow: siRNA-1 5′-UAA​ACA​CCG​UCA​CUU​CUC​CTT-3′, siRNA-2 5′-UAA​UAA​GGC​ACU​UCC​AUC​CTT-3′, siRNA-3 5′-AUG​UUC​UUC​AUG​AUC​UCC​GTT-3’.

E10.5 Embryos were fixed in 4% Paraformaldehyde at 4°C overnight, equilibrated in 1xPBS/15% sucrose and 1xPBS/30% sucrose sequentially till they sink to the bottom and then were embedded in OCT. Embedded embryos were sectioned with 20 µm thickness per section. Embryo section slides were washed in 1xPBS solution to remove OCT, then fixed in 4% Paraformaldehyde for 10 min, washed in 1xPBS for 3 × 3 min and permeabilized in 1xPBS/0.5% Triton-X 100 at 4°C for 2 h. Permeabilized sections were blocked in 1xPBS/0.05% Triton-X 100 with 5% goat serum for 2 h, incubated in primary antibodies at 4°C overnight, incubated in secondary antibodies avoiding light at room temperature for 2 h (antibodies listed in Supplementary Table S5). Then sections were immersed in 1xPBS/DAPI for 10 min, rinsed in 1xPBS 3 × 3 min and sealed for further imagingImmunofluorescence images of embryonic sections were taken with a Zeiss LSM 700 confocal microscope.

For immunoblotting, E10.5 mouse embryo tissues were lysed in pre-chilled 1% Triton X-100 buffer (20 mM Tris buffer at pH7.5, with 150 mM NaCl plus sodium pyrophosphate, β-glycerophosphate, EDTA, Na₃VO₄, and leupeptin) for 10 min on ice. Protein lysates were collected after high-speed centrifugation and mixed with 5x SDS-PAGE loading buffer and denatured at 95°C for 5 min. Denatured proteins were separated by SDS-PAGE, electro-transferred to PVDF membrane and probed with primary and secondary antibodies (antibodies used in Supplementary Table S5). Blot images were captured on Tanon™ 5200CE Chemi-Image System and analyzed in ImageJ.

For immunoprecipitation, protein lysates from 293T cells were prepared as above and incubated in 1xPBS/IP antibodies on rotator at 4°C overnight. Then protein solutions were incubation with antibody and Protein A + G Agarose beads on rotator at 4°C for 2 h. Agarose beads were washed in cold 1xPBS and collected by 2,500 rpm centrifugation for three times. Proteins were eluted from beads with 5x SDS-PAGE loading buffer and denatured at 95°C for 5 min. Denatured immunoprecipitants were acquired by high-speed centrifugation, and analyzed by electrophoresis as described in immunoblotting.

Protein lysates were gained and separated by SDS-PAGE as described above. Whole SDS-PAGE gel was stained by Coomassie brilliant blue. Stained gel zones parallel to 100 kDa and 75 kDa markers were cut out for Mass spectrometry analysis. Mass spectrometry experiments were performed at the State Key Laboratory of Genetic Engineering Fudan University.