Introductory paragraph
Multiciliated cells (MCCs) harbour dozens to hundreds of motile cilia, which beat in a synchronized and directional manner, thus generating hydrodynamic forces important in animal physiology1, 2. In vertebrates, MCC differentiation critically depends on the synthesis and release of numerous centrioles by specialized structures called deuterosomes1-5. Little is known about the composition, organization and regulation of deuterosomes. Here, single-cell RNA sequencing reveals that human deuterosome-stage MCCs are characterized by the expression of many cell cycle-related genes. Among those, we further investigated the uncharacterized vertebrate-specific cell division cycle 20B (CDC20B) gene. We show that the CDC20B protein associates to the deuterosome, and is required for the production of centrioles and cilia in mouse and Xenopus MCCs. In Xenopus, centrioles and cilia were efficiently rescued in absence of CDC20B by over-expression of the protease Separase, linking CDC20B function to centriole release from deuterosomes, in analogy to centriole disengagement in mitotic cells. This work reveals the shaping of a new biological function, deuterosome-mediated centriole production in vertebrate MCCs, by adaptation between ancestral and recently evolved cell cycle-related molecules.
Main text
Multiciliated cells (MCCs) are present throughout metazoan evolution and serve functions ranging from locomotion of marine larvae and flatworms, to brain homeostasis, mucociliary clearance of pathogens and transportation of oocytes in vertebrates1, 2, 6. The formation of MCCs requires the production of numerous motile cilia through a complex process called multiciliogenesis1, 2. Recently, defective multiciliogenesis has been associated with human congenital respiratory and fertility syndromes7, 8. Each cilium sits atop a modified centriole, called a basal body (BB). After they exit from the cell cycle, maturing MCCs face the challenge of producing hundreds of centrioles in a limited time window. In vertebrate MCCs, bulk centriole biogenesis is mostly achieved through an acentriolar structure named the deuterosome, although canonical amplification from parental centrioles also occurs1, 2, 6. Recent studies have suggested that deuterosome-mediated centriole synthesis mobilizes proteins usually involved in the centriole-dependent duplication pathway of the cell cycle, such as PLK4, CEP152 and SAS63-5. However, the deuterosome pathway also mobilizes specific components, structurally related to cell cycle molecules, such as the Geminin-related protein Multicilin (MCIDAS in mammals), which is necessary and sufficient to drive multiciliogenesis9, or the CEP63 paralogue DEUP1, an essential component of deuterosome assembly5. To better understand deuterosome biology, we used single-cell RNA sequencing (scRNA-seq) to identify genes expressed in maturing MCCs during the phase of centriole amplification.
We applied scRNA-seq analyses to 3D air-liquid primary cultures of human airway epithelial cells (HAECs) at the differentiation stage corresponding to active centriole multiplication10 (Fig. 1a). Gene expression data from 1663 cells was projected on a 2D space by t-distributed Stochastic Neighbor Embedding (tSNE) (Fig. 1b). We identified a small group of 37 cells corresponding to maturing MCCs engaged in deuterosome-mediated centriole amplification, as revealed by the specific expression of MCIDAS9, MYB11, and DEUP15 (Fig. 1c,d and Supplementary Fig. 1). This subpopulation was characterized by the expression of known effectors of centriole synthesis, such as PLK4, STIL, CEP152, SASS6, but also of cell cycle regulators, such as CDK1, CCNB1, CDC20, SGOL2 and NEK2 (Fig. 1d and Supplementary Fig. 1). We reasoned that uncharacterized cell cycle-related genes that are specific to this subpopulation could encode new components of the deuterosome-dependent centriole amplification pathway. A particularly interesting candidate in this category was CDC20B (Fig. 1d), which is related to the cell cycle regulators CDC20 and FZR112 (Supplementary Fig. 2a). First, the CDC20B gene is present in the vertebrate genomic locus that also contains the key MCC regulators MCIDAS9 and CCNO8. Co-expression of CDC20B, MCIDAS and CCNO throughout HAEC differentiation was indeed observed in an independent RNA sequencing study, performed on a bulk population of HAECs (Supplementary Fig. 2c). These results fit well with the observation that the promoter of human CDC20B was strongly activated by the MCIDAS partners E2F1 and E2F4 (Supplementary Fig. 2d), as also shown in Xenopus by others13 (Supplementary Fig. 2e). Second, the CDC20B gene bears in its second intron the miR-449 microRNAs (Supplementary Fig. 3k), which were shown to contribute to MCC differentiation10, 14-17. Finally, in Xenopus epidermal MCCs, cdc20b transcripts were specifically detected during the phase of centriole amplification (Supplementary Fig. 3a-i). This first set of data pointed out the specific and conserved expression pattern of CDC20B in immature MCCs, and prompted us to analyze its putative role in centriole amplification.
We then analyzed the subcellular localization of CDC20B protein in human, mouse and Xenopus MCCs. As previously reported, the CDC20B protein was detected near BBs10, but also in cilia of fully differentiated human airway MCCs (Supplementary Fig. 2f-h). This was confirmed by proximity ligation assays that revealed a tight association of CDC20B with Centrin2 and acetylated α-tubulin, in BBs and cilia, respectively (Supplementary Fig. 2i-k). In immature mouse tracheal MCCs, double immunofluorescence also revealed the association of CDC20B to Deup1-positive deuterosomes (Fig. 2a and Supplementary Fig. 2b). Likewise, CDC20B decorated deuterosomes, which were revealed by the centriole marker FOP in immature mouse ependymal MCCs (Fig. 2b). Finally, we found that RFP-Cdc20b localized around GFP-Deup1 positive structures in immature Xenopus epidermal MCCs (Fig. 2c). We conclude that in vertebrate MCCs, CDC20B is associated to deuterosomes during the phase of centriole production.
Next, Cdc20b was knocked down in mouse ependymal MCCs, through post-natal brain electroporation of three distinct shRNAs (Supplementary Fig. 3j). One of them, sh274, which targets the junction between exons 3 and 4, and can therefore only interact with mature mRNA, was useful to rule out possible interference with the production of miR-449 molecules from the CDC20B pre-mRNA (Supplementary Fig. 3j). Despite the reduced expression of CDC20B by all three shRNAs (Fig. 3c), MCC identity was not affected as revealed by FOXJ1 expression (Fig. 3a,b,d). In sharp contrast to control shRNA (Fig. 3e), Cdc20b shRNAs caused 50% to 70% of abnormal deuterosomal figures in early stage MCCs, including stalled, fused, oversized or minute deuterosomes (Fig. 3f-h). Late stage CDC20B-deficient ependymal MCCs did not undergo proper multiciliogenesis, and often displayed few centrioles and aberrant deuterosomal figures, suggesting a blockage in the centriole biosynthesis pathway (Fig. 3i-k).
Cdc20b was also knocked down in Xenopus epidermal MCCs, through injection of two independent morpholino antisense oligonucleotides targeting either the ATG (Mo ATG), or the exon 1/intron 1 junction (Mo Spl) (Supplementary Fig. 3k). The efficiency of Mo ATG was verified through fluorescence extinction of co-injected cdc20b-GFP mRNA (Supplementary Fig. 3m). RT-PCR confirmed that Mo Spl caused intron 1 retention (Supplementary Fig. 3l), which is expected to introduce a premature stop codon, and to produce a Cdc20b protein lacking 96% of its amino-acids, likely to undergo unfolded protein response-mediated degradation. Thus, both morpholinos were expected to generate severe loss of Cdc20b function. We verified that neither morpholinos caused p53 transcript up-regulation (Supplementary Fig. 3n), a non-specific response to morpholinos that is sometimes detected in zebrafish embryos18. Importantly, whole-mount in situ hybridization indicated that miR-449 expression was not perturbed in the presence of either morpholinos (Supplementary Fig. 3o). We found that cdc20b knockdown did not interfere with acquisition of the MCC fate (Supplementary Fig. 4a-e), but severely impaired multiciliogenesis, as revealed by immunofluorescence and electron microscopy (Fig. 4a-i). This defect stemmed from a dramatic reduction in the number of centrioles, and poor docking at the plasma membrane (Fig. 4g-o and Supplementary Fig. 4f-k). Late-stage morphant MCCs that lacked cilia often displayed stalled deuterosomal figures, again suggesting a blockage in the centriole biosynthesis pathway (Fig. 5l and Supplementary Fig. 4h). Importantly, centrioles and cilia were rescued in Mo Spl MCCs by co-injection of cdc20b, venus-cdc20b or cdc20b-venus mRNAs (Fig. 4j-o and Supplementary Fig. 4f-k). In normal condition, Xenopus epidermal MCCs arise in the inner mesenchymal layer and intercalate into the outer epithelial layer, while the process of centriole amplification is underway. To rule out secondary defects due to poor radial intercalation, we assessed the consequences of cdc20b knockdown in MCCs induced in the outer layer by Multicilin overexpression9. Like in natural MCCs, Cdc20b proved to be essential for the production of centrioles and cilia in response to Multicilin activity (Supplementary Fig. 5a-g). We also noted that the apical actin network that normally surrounds BBs was not maintained in absence of Cdc20b, although this defect could be secondary to the absence of centrioles (Supplementary Fig. 5d-g). Altogether our functional assays indicate that CDC20B is required for the production of multiple centrioles in vertebrate MCCs.
CDC20B encodes a protein of about 519 amino-acids largely distributed across the vertebrate phylum10. In its C-terminal half, CDC20B contains seven well conserved WD40 repeats, predicted to form a β-propeller, showing 49% and 37% identity to CDC20 and FZR1 repeats, respectively (Supplementary Fig. 2a). While CDC20 and FZR1 bind and activate the Anaphase Promoting Complex (APC/C), CDC20B lacks canonical APC/C binding domains (Supplementary Fig. 2a), suggesting a distinct molecular function. Insight into this function was suggested by an unbiased interactome study, which reported an association of CDC20B with the mitotic kinase PLK119. Phosphorylation of the pericentriolar material protein Pericentrin (PCNT) by PLK1 induces its cleavage by the protease Separase, thereby allowing centriole disengagement at the end of mitosis20-22. Based on our phenotypic observations, we hypothesized that the PLK1/Separase/PCNT axis could not only control centriole disengagement in mitotic cells, but also centriole release from deuterosomes in MCCs (Fig. 5o). Consistent with this idea, PLK1, Separase (ESPL1), its inhibitor Securin (PTTG1), and PCNT were found to be expressed in human deuterosome-stage MCCs (Supplementary Fig. 1), and/or in Xenopus epidermal MCCs13. The corresponding proteins were all detected in maturing murine ependymal MCCs (Fig. 5a-f). Of particular relevance, PCNT showed a prominent association to deuterosomes actively engaged in centriole synthesis (Fig. 5f). STED super-resolution fluorescent microscopy further revealed that PCNT was tightly associated to centriolar walls, consistent with a role in the physical maintenance of centrioles around deuterosomes (Fig. 5g). We also note that PLK1 was found enriched around deuterosomes of cultured murine ependymal MCCs prior to centriole disengagement23. We functionally tested our hypothesis, by over-expressing human Separase in Xenopus cdc20b morphant MCCs to force the release of centrioles from deuterosomes. Over-expression of wild-type, but not protease-dead Separase, efficiently rescued the production of centrioles and cilia in Cdc20b-deficient MCCs (Fig. 5h-n and Supplementary Fig. 5k-p). Separase could also rescue multiciliogenesis in Multicilin-induced MCCs injected with cdc20b morpholinos (Supplementary Fig. 5q-w). Interestingly, we noticed that overexpression of Cdc20b or Separase both caused the formation of multipolar spindles in non-MCC dividing cells, consistent with forced centriole disengagement24 (Supplementary Fig. 5h-j). These results suggest that the function of CDC20B in early-stage MCCs consists of allowing the release by Separase of mature centrioles from deuterosomes.
In this study, we report for the first time the essential and conserved role of CDC20B in vertebrate multiciliogenesis. Our data suggest that the presence of CDC20B around deuterosomes is necessary to allow Separase-dependent proteolysis leading to centriole disengagement. How Separase is activated in maturing MCCs remains to be addressed. During the cell cycle, Separase is activated through APC/C-mediated degradation of its inhibitor Securin, which involves activation of APC/C by CDC20 or FZR112. We found that CDC20 and FZR1 are expressed in human deuterosome-stage MCCs (Fig. 1d and Supplementary Fig. 1). CDC20 protein was detected in cultured murine ependymal MCCs during the phase of centriole disengagement23, and FZR1 genetic ablation was reported to cause reduced production of centrioles and cilia in mouse ependymal MCCs25. Although not formally excluded, direct APC/C activation by CDC20B is unlikely, based on its structure (Supplementary Fig. 2a) and on unbiased interactome studies, which failed to recover interaction with any of the APC/C complex components19, 26. APC/C is therefore likely activated in maturing MCCs by its classical partners, CDC20 and/or FZR1, leading to Separase activation through degradation of its inhibitor Securin. Additional factors may contribute to activation of Separase. For instance, we noticed that the APC/C inhibitor FBXO43 (EMI2)27 is strongly up-regulated in mouse and frog MCCs13, 28. As PLK1 is known to induce SCF-mediated degradation of EMI229, it could coordinate Separase activation in the cytoplasm and PCNT priming on deuterosomes. Also of interest, SPAG5 (Astrin), a common interactor of DEUP1 and CDC20B26, 30, was reported to control timely activation of Separase during the cell cycle24. It is therefore possible that multiple modes of activation of Separase act in parallel to trigger the release of neo-synthesized centrioles in maturing MCCs. Alternatively, different pathways may be used in distinct species, or in distinct types of MCCs. We found that beyond its association to deuterosomes during the phase of centriole amplification, CDC20B was also associated to basal bodies and cilia in fully differentiated MCCs. This dual localization in mature MCCs is consistent with poor BB apical docking and cilium growth caused by CDC20B knockdown in our models. However, refined temporal and spatial control of CDC20B inhibition will be needed to study its putative role beyond centriole synthesis.
This and previous studies10, 14-17 establish that the miR-449 cluster and its host gene CDC20B are commonly involved in multiciliogenesis. Consistent with its early expression, it was suggested that miR-449 controls cell cycle exit and entry into differentiation of MCCs10, 15. This study reveals that CDC20B is involved in the production of centrioles, the first key step of the multiciliogenesis process. From that perspective, the nested organization of miR-449 and CDC20B in vertebrate genomes, which allows their coordinated expression, appears crucial for successful multiciliogenesis.
Recently, congenital mutations in MCIDAS and CCNO were shown to cause a newly-recognized MCC-specific disease, called Reduced Generation of Multiple motile Cilia (RGMC), characterized by severe chronic lung infections and increased risk of infertility7, 8. Its location in the same genetic locus as MCIDAS and CCNO makes CDC20B a good candidate for RGMC. By extension, new deuterosome-stage specific genes uncovered by scRNA-seq in this study also represent potential candidates for additional RGMC mutations. Previous works have established the involvement of the centriole duplication machinery active in S-phase of the cell cycle, during centriole multiplication of vertebrate post-mitotic MCCs3-5. Our study further reveals a striking analogy between centriole disengagement from deuterosomes in MCCs, and centriole disengagement that occurs during the M/G1 transition of the cell cycle (Fig. 5o). Thus, it appears that centriole production in MCCs recapitulates the key steps of the centriole duplication cycle31. However, the cell cycle machinery must adapt to the acentriolar deuterosome to massively produce centrioles. Such adaptation appears to involve physical and functional interactions between canonical cell cycle molecules, such as CEP152 and PLK1, and recently evolved cell cycle-related deuterosomal molecules, such as DEUP15 and CDC20B. It remains to examine whether additional deuterosomal molecules have emerged in the vertebrate phylum to sustain massive centriole production in MCCs.
In conclusion, this work illustrates how coordination between ancestral and recently evolved cell cycle-related molecules can give rise to a new differentiation mechanism in vertebrates.
Author contributions
PB, BM and LK designed and supervised the study, and obtained funding. LEZ, SRG, OM performed and analyzed human and mouse airway cells experiments. DRR and VT performed and analyzed Xenopus experiments. CB performed and analyzed experiments on mouse ependymal MCCs. OR performed WB analysis. MD and AP performed the bioinformatic analysis. All authors were involved in data interpretation. DRR, LEZ and CB designed the figures. LK, PB, DRR, LEZ, CB, and BM wrote the manuscript.
Competing financial interests
The authors declare no competing financial interests.
Materials and Methods
Subjects/human samples
Inferior turbinates were from patients who underwent surgical intervention for nasal obstruction or septoplasty (provided by L. Castillo, Nice University Hospital, France). The use of human tissues was authorized by the bioethical law 94-654 of the French Public Health Code and written consent from the patients.
Single-cell RNA sequencing of HAECs
HAECs were cultured as previously described 10. They were induced to differentiate at the air-liquid interface for 14 days, which corresponds to the maximum centriole multiplication stage. Cells were incubated on Transwell® (Corning®, NY 14831 USA) with 0.1% protease type XIV from Streptomyces griseus (Sigma-Aldrich) in HBSS (Hanks' balanced salts) for 4 hours at 4°C degrees. Cells were gently detached from the Transwells by pipetting and then transferred to a microtube. 50 units of DNase I (EN0523 Thermo Fisher Scientific) per 250µl, were directly added and cells were further incubated at room temperature for 10 min. Cells were centrifuged (150g for 5 min) and resuspended in 500 µL HBSS 10% Fetal Bovine Serum (Gibco), centrifuged again (150g for 5 min) and resuspended in 500 µL HBSS before being mechanically dissociated through a 26G syringe (4 times). Finally, cell suspensions were filtered through a Scienceware® Flowmi™ Cell Strainer (40µm porosity), centrifuged (150g for 5 min) and resuspended in 500 µL of cold HBSS. Cell concentration measurements were performed with Scepter™ 2.0 Cell Counter (millipore) and Countess™ automated cell counter (ThermoFisher Scientific). Cell viability was checked with Countess™ automated cell counter (ThermoFisher Scientific). All steps except the DNAse I incubation were performed on ice. For the cell capture by the 10X genomics device, the cell concentration was adjusted to 300 cells/µl in HBSS aiming to capture 1500 cells. We then followed the manufacturer’s protocol (Chromium™ Single Cell 3' Reagent Kit, v2 Chemistry) to obtain single cell 3’ libraries for Illumina sequencing. Libraries were sequenced with a NextSeq 500/550 High Output v2 kit (75 cycles) that allows up to 91 cycles of paired-end sequencing: the forward read had a length of 26 basse that included the cell barcode and the UMI; the reverse read had a length of 57 bases that contained the cDNA insert. CellRanger Single-Cell Software Suite v1.3 was used to perform sample demultiplexing, barcode processing and single-cell 3′ gene counting using standards default parameters and human build hg19. Additional analyses were performed using R. Pseudotemporal ordering of single cells was performed with the last release of the Monocle package 32. Cell cycle scores were calculated by summing the normalized intensities of genes belonging to phase-specific gene sets then centered and scaled by phase. Gene sets for each phase were curated from previously described sets of genes 33(Table S1). Data was submitted to the GEO portal under series reference GSE103518. Data shown in Figure 1 is representative of 4 independent experiments performed on distinct primary cultures.
RNA sequencing of HAECs
For Figure S3A, three independent HAEC cultures (HAEC1, HAEC2, HAEC3) were triggered to differentiate in air-liquid interface (ALI) cultures for 2 days (ALI day 2, undifferentiated), ALI day 14 (first cilia) or ALI day 28 (well ciliated). RNA was extracted with the miRNeasy mini kit (Qiagen) following manufacturer’s instructions. mRNA-seq was performed from 2 µg of RNA that was first subjected to mRNA selection with Dynabeads® mRNA Purification Kit (Invitrogen). mRNA was fragmented 10 min at 95°C in RNAseIII buffer (Invitrogen) then adapter-ligated, reverse transcribed and amplified (6 cycles) with the reagents from the NEBNext Small RNA Library Prep Set for SOLiD. Small RNA-seq was performed from 500 ng RNA with the NEBNext Small RNA Library Prep Set for SOLiD (12 PCR cycles) according to manufacturer’s instructions. Both types of amplified libraries were purified on Purelink PCR micro kit (Invitrogen), then subjected to additional PCR rounds (8 cycles for RNA-seq and 4 cycles for small RNA-seq) with primers from the 5500 W Conversion Primers Kit (Life Technologies). After Agencourt® AMPure® XP beads purification (Beckman Coulter), libraries were size-selected from 150 nt to 250 nt (for RNA-seq) and 105 nt to 130 nt (for small RNA-seq) with the LabChip XT DNA 300 Assay Kit (Caliper Lifesciences), and finally quantified with the Bioanalyzer High Sensitivity DNA Kit (Agilent). Libraries were sequenced on SOLiD 5500XL (Life Technologies) with single-end 50b reads. SOLiD data were analyzed with lifescope v2.5.1, using the small RNA pipeline for miRNA libraries and whole transcriptome pipeline for RNA-seq libraries with default parameters. Annotation files used for production of raw count tables correspond to Refseq Gene model v20130707 for mRNAs and miRBase v18 for small RNAs. Data generated from RNA sequencing were then analyzed with Bioconductor (http://www.bioconductor.org) package DESeq and size-factor normalization was applied to the count tables. Heatmaps were generated with GenePattern using the “Hierarchical Clustering” Module, applying median row centering and Euclidian distance.
Re-analysis of Xenopus E2F4 Chip-seq and RNA-seq
RNA-seq (samples GSM1434783 to GSM1434788) and ChIP-seq (samples GSM1434789 to GSM1434792) data were downloaded from GSE59309. Reads from RNA-seq were aligned to the Xenopus laevis genome release 7.1 using TopHat2 34 with default parameters. Quantification of genes was then performed using HTSeq-count 35 release 0.6.1 with “-m intersection-nonempty” option. Normalization and statistical analysis were performed using Bioconductor package DESeq2 36. Differential expression analysis was done between Multicilin-hGR alone versus Multicilin-hGR in the presence of E2f4ΔCT. Reads from ChIP seq were mapped to the Xenopus laevis genome release 7.1 using Bowtie2 37. Peaks were called and annotated according to their positions on known exons with HOMER 38. Peak enrichments of E2F4 binding site in the promoters of centriole genes and cell cycle genes 13 were estimated in presence or absence of Multicilin and a ratio of E2F4 binding (Multicilin vs no Multicilin) was calculated.
Promoter reporter studies
The human CDC20B promoter was cloned into the pGL3 Firefly Luciferase reporter vector (Promega) with SacI and NheI cloning sites. The promoter sequenced ranged from −1073 to +104 relative to the transcription start site. 37.5 ng of pGL3 plasmid were applied per well. pCMV6-Neg, pCMV6-E2F1 (NM_005225) and pCMV6-E2F4 (NM_001950) constructs were from Origene. 37.5 ng of each plasmid was applied per well. 25 ng per well of pRLCMV (Promega) was applied in the transfection mix for transfection normalization (Renilla luciferase). HEK 293T cells were seeded at 20 000 cells per well on 96-well plates. The following day, cells were transfected with the indicated plasmids (100 ng of total DNA) with lipofectamine 3000 (Invitrogen). After 24 hours, cells were processed with the DualGlo kit (Promega) and luciferase activity was recorded on a plate reader.
Proximity ligation Assays
Fully differentiated HAECs were dissociated by incubation with 0.1% protease type XIV from Streptomyces griseus (Sigma-Aldrich) in HBSS (Hanks' balanced salts) for 4 hours at 4°C. Cells were gently detached from the Transwells by pipetting and then transferred to a microtube. Cells were then cytocentrifuged at 300 rpm for 8 min onto SuperFrostPlus slides using a Shandon Cytospin 3 cytocentrifuge. Slides were fixed for 10 min in methanol at −20°C for Centrin2 and ZO1 assays, and for 10 min in 4% paraformaldehyde at room temperature and then permeabilized with 0.5% Triton X-100 in PBS for 10 min for Acetylated-α-tubulin assays. Cells were blocked with 3% BSA in PBS for 30 min. The incubation with primary antibodies was carried out at room temperature for 2 h. Then, mouse and rabbit secondary antibodies from the Duolink® Red kit (Sigma-Aldrich) were applied and slides were processed according to manufacturer’s instructions. Images were acquired using the Olympus Fv10i confocal imaging systems with 60X oil immersion objective and Alexa 647 detection parameters.
Mice
Timed pregnant CD1 mice were used (Charles Rivers, Lyon, France). Animal experiments were carried out in accordance to European Community Council Directive and approved by French ethical committees (comité d’éthique pour l’expérimentation animale n°14; permission number: 62-12112012).
Immunostaining on mouse ependyma
Immunostaining on ependyma preparations were performed as previously described 39. Briefly, dissected brains were subjected to 12 min fixation in 4% paraformaldehyde, 0.1% Triton X-100, blocked 1 hour in PBS, 3% BSA, incubated overnight with primary antibodies diluted in PBS, 3% BSA, and incubated 1 h with secondary antibodies at room temperature. Ependyma were dissected further and mounted with Mowiol before imaging using an SP8 confocal microscope (Leica microsystems) equipped with a 63x oil objective. The same protocol was used to prepare samples for super-resolution acquisition. Pictures were acquired with a TCS SP8 STED 3X microscope equipped with an HC PL APO 93X/1.30 GLYC motCORRTM objective (Leica microsystems). Pericentrin was revealed using Alexa 514 (detection 535-564nm, depletion 660nm) and FOP was revealed using Alexa 488 (detection 498-531nm, depletion 592nm). Pictures were deconvoluted using Huygens software. Maximum intensity projection of 3 deconvoluted pictures is presented in Figure 4G. Primary antibodies: rabbit anti-CDC20B (1:500; Proteintech), mouse IgG anti-PLK1 (1:500; Thermo Fisher), rabbit anti-Pericentrin (1:500, Abcam), mouse IgG2a anti-Securin (1:100; Abcam), rabbit anti-Separase (1:200; Abcam), mouse IgG1 anti-FoxJ1 (1:1000; eBioscience), rabbit anti-Deup1 (1:1000; kindly provided by Dr Xueliang Zhu), rabbit anti-ZO1 (1:600; Thermo Fisher Scientific), mouse IgG1 anti-ZO1 (1:600; Invitrogen), mouse IgG2b anti-FGFR1OP (FOP) (1:2000; Abnova), mouse IgG1 anti-α-tubulin (1:500; Sigma-Aldrich). Secondary antibodies: Alexa Fluor 488 goat anti-rabbit (1:800; Thermo Fisher Scientific), Alexa Fluor 647 goat anti-rabbit (1:800; Thermo Fisher Scientific), Alexa Fluor 514 goat anti-rabbit (1:800; Thermo Fisher Scientific), Alexa Fluor 488 goat anti-mouse IgG2b (1:800; Thermo Fisher Scientific), Alexa Fluor 568 goat anti-mouse IgG2b (1:800; Thermo Fisher Scientific), Alexa Fluor 488 goat anti-mouse IgG2a (1:800; Thermo Fisher Scientific), Alexa Fluor 568 goat anti-mouse IgG1 (1:800; Thermo Fisher Scientific), Alexa Fluor 647 goat anti-mouse IgG1 (1:800; Thermo Fisher Scientific).
Mouse constructs
Expression constructs containing shRNA targeting specific sequences in the CDC20B coding sequence under the control of the U6 promoter were obtained from Sigma-Aldrich (ref. TRCN0000088273 (sh273), TRCN0000088274 (sh274), TRCN0000088277 (sh277)). PCX-mcs2-GFP vector (Control GFP) kindly provided by Xavier Morin (ENS, Paris, France), and U6 vector containing a validated shRNA targeting a specific sequence in the NeuroD1 coding sequence 40(Control sh, ref. TRCN0000081777, Sigma-Aldrich) were used as controls for electroporation experiments.
Postnatal mouse brain electroporation
Postnatal mouse brain electroporation was performed as described previously 41. Briefly, P1 pups were anesthetized by hypothermia. A glass micropipette was inserted into the lateral ventricle, and 2 µl of plasmid solution (concentration 3μg/μl) was injected by expiratory pressure using an aspirator tube assembly (Drummond). Successfully injected animals were subjected to five 95V electrical pulses (50 ms, separated by 950 ms intervals) using the CUY21 edit device (Nepagene, Chiba, Japan), and 10 mm tweezer electrodes (CUY650P10, Nepagene) coated with conductive gel (Signagel, Parker laboratories). Electroporated animals were reanimated in a 37°C incubator before returning to the mother.
Statistical analyses of mouse experiments
Analysis of Cdc20b shRNAs efficiency (Fig. 3c): For each field, the intensity of CDC20B fluorescent immunostaining was recorded using ImageJ software and expressed as arbitrary units. Data are mean ± sem. Two independent experiments were analyzed. A minimum of 39 fields per condition was analyzed. n= 3, 4, 5 and 5 animals for sh control, sh273, sh274 and sh277, respectively. Unpaired t test vs sh control: p=0.0036 (sh273, **), 0.0135 (sh274, *), 0.0035 (sh277, **).
Analysis of the number of FOXJ1 positive cells at 5dpe (Fig. 3d): Unpaired t test vs sh control: 0.3961 (sh273, ns), 0.1265 (sh274, ns), 0.3250 (sh277, ns).
Analysis of the number of deuterosomal figures at 5dpe (Fig. 3g): 50-100 cells with deuterosomal figures were analyzed per condition. n= 3 4, 2, and 3 animals for sh control, sh273, sh274 and sh277, respectively. Unpaired t test vs sh control: p=0.0059 (sh273, **), 0.4091 (sh274, ns), 0.4805 (sh277, ns). sh273 caused significant increase in the number of deuterosomal figures.
Analysis of deuterosome categories at 5dpe (Fig. 3h): Data are mean ± sem from two independent experiments. 50-100 cells were analyzed for each condition. Unpaired t test vs sh control: p=0.0008 (sh273, ***), 0.0164 (sh274, *), 0.0158 (sh277, *).
Analysis of ependymal cells categories at 15dpe (Fig. 3k): Data are mean ± sem from three independent experiments. More than 500 cells were analyzed for each condition. n= 4, 4, 3, and 3 animals for sh control, sh273, sh274 and sh277 respectively. Unpaired t test vs sh control: p= 0.0004 (sh273, ***), 0.0001 (sh274, ****), 0.0038 (sh277, **).
Mouse tracheal epithelial cells (MTECs)
MTECs cell cultures were established from the tracheas of 12 weeks old mice, according to the procedure previously published 42, with the following modification: in differentiation medium, NuSerumTM was replaced with Ultroser-GTM (Pall Corporation) and 10 µM DAPT (N-[N-(3,5-difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester) (Sigma) was added one day after setting-up the air-liquid interface.
Immunostaining on HAECs and MTECs
Three days after setting-up the air-liquid interface, MTECs on Transwell membranes were pre-extracted with 0.5% Triton X-100 in PBS for 3 min, and then fixed with 4% paraformaldehyde in PBS for 15 min at room temperature. HAECs were treated 21 days after setting-up the air-liquid interface. They were fixed directly on Transwells with 100% cold methanol for 10 min at −20°C (for CDC20B and Centrin2 co-staining, Fig. S5A,B) or with 4% paraformaldehyde in PBS for 15 min at room temperature (for CDC20B single staining, Fig. S5C). All cells were then permeabilized with 0.5% Triton X-100 in PBS for 10 min and blocked with 3% BSA in PBS for 30 min. The incubation with primary and secondary antibodies was carried out at room temperature for 2 h and 1 h, respectively. Nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI). Transwell membranes were cut with a razor blade and mounted with ProLong Gold medium (Thermofisher). Primary antibodies: rabbit anti-CDC20B (1:500; Proteintech), rabbit anti-DEUP1 (1:500), anti-centrin2 (Clone 20H5, 1:500; Millipore). Secondary antibodies: Alexa Fluor 488 goat anti-rabbit (1:1000; Thermo Fisher Scientific), Alexa Fluor 647 goat anti-mouse (1:1000; Thermo Fisher Scientific). For co-staining of CDC20B and DEUP1, CDC20B was directly coupled to CFTM 594 with the Mix-n-StainTM kit (Sigma-Aldrich) according to the manufacturer’s instruction. Coupled primary antibodies were applied after secondary antibodies had been extensively washed.
Cells, transfection and western blot analysis
Cos1 cells were grown in DMEM supplemented with 10% heat inactivated FCS and transfected with Fugene HD (Roche Applied Science) according to manufacturer’s protocol. Transfected or control cells were washed in PBS and lysed in 50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1mM EDTA, containing 1% NP-40 and 0.25% sodium deoxycholate (modified RIPA) plus a Complete Protease Inhibitor Cocktail (Roche Applied Science) on ice. Cell extracts separated on polyacrylamide gels were transfered onto Optitran membrane (Whatman) followed by incubation with rabbit anti-CDC20B antibody (Proteintech) and horseradish peroxidase conjugated secondary antibody (Jackson Immunoresearch Laboratories). Signal obtained from enhanced chemiluminescence (Western Lightning ECL Pro, Perkin Elmer) was detected with MyECL Imager (Thermo).
Xenopus embryo injections, plasmids, RNAs, and Mos
Eggs obtained from NASCO females were fertilized in vitro, dejellied and cultured as described previously 43. All injections were done at the 8-cell stage in one animal-ventral blastomere (presumptive epidermis), except for electron microscopy analysis for which both sides of the embryo was injected, and for RT-PCR analysis for which 2-cell embryos were injected.
cdc20b riboprobe was generated from Xenopus laevis cDNA. Full-length sequence was subcloned in pGEM™-T Easy Vector Systems (Promega). For sense probe it was linearized by SpeI and transcribed by T7. For antisense probe it was linearized by ApaI and transcribed by Sp6 RNA polymerase. Synthetic capped mRNAs were produced with the Ambion mMESSAGE mMACHINE Kit. pCS105/mGFP-CAAX was linearized with AseI and mRNA was synthesized with Sp6 polymerase. pCS2-mRFP and pCS2-GFPgpi were linearized with NotI and mRNA was synthesized with Sp6 polymerase. pCS-centrin4-YFP (a gift from Reinhard Köster, Technische Universität Braunschweig, Germany) was linearized with Notl and mRNA was synthesized with Sp6 polymerase. pCS2-GFP-Deup1 and pCS2-Multicilin(MCI)-hGR were kindly provided by Chris Kintner and the mRNAs were obtained as described previously 9. Embryos injected with MCI-hGR mRNA were cultured in Dexamethasone 20μM in MBS 0,1X from st11 until fixation. pCS2-Separase wild-type and phosphomutant 2/4 (protease dead, PD) were provided by Marc Kirchner and Olaf Stemann, respectively; plasmids were linearized with NotI and mRNAs were synthesized with Sp6 polymerase. Venus-cdc20b, cdc20b-Venus and cdc20b were generated by GATEWAY™ Cloning Technology (GIBCO BRL) from Xenopus laevis cdc20b cDNA. cdc20b was also subcloned in pCS2-RFP to make RFP-cdc20b and cdc20b-RFP fusions. All cdc20b constructs were linearized with NotI and mRNAs were synthesized with Sp6 polymerase. Quantities of mRNA injected: 0.5ng for GFP-CAAX, RFP, GFP-GPI GFP-Deup1, MCI-hGR, Separase and Separase(PD); 1ng for Venus-cdc20b, cdc20b-Venus, cdc20b, RFP-cdc20b and cdc20b-RFP.
Two independent morpholino antisense oligonucleotides were designed against cdc20b (GeneTools, LLC). cdc20b ATG Mo: 5'-aaatcttctctaacttccagtccat-3', cdc20b Spl Mo 5'-acacatggcacaacgtacccacatc-3'. 20ng of MOs was injected per blastomere or 10ng of each Mo for co-injection.
PCR and Quantitative RT-qPCR
Xenopus embryos were snap frozen at different stages and stored at −80°C. Total RNAs were purified with a Qiagen RNeasy kit (Qiagen).
Primers were designed using Primer-BLAST Software. PCR reactions were carried out using GoTaq® G2 Flexi DNA Polymerase (Promega). RT reactions were carried out using iScript™ Reverse Transcription Supermix for RT-qPCR (BIO-RAD). qPCR reactions were carried out using SYBRGreen on a CFX Biorad qPCR cycler. To check cdc20b temporal expression by qPCR we directed primers to exons 9/10 junction (Forward: 5'-ggctatgaattggtgcccg-3') and exons 10/11 junction (Reverse: 5'-gcagggagcagatctggg-3') to avoid amplification from genomic DNA. The relative expression of cdc20b was normalized to the expression of the housekeeping gene ornithine decarboxylase (ODC) for which primers were as follows: forward: 5'-gccattgtgaagactctctccattc-3': reverse: 5'-ttcgggtgattccttgccac-3'. To check the efficiency of Mo SPL, expected to cause retention of intron1 in the mature mRNA of cdc20b we directed forward (5'-cctcccgagagttagagga-3') and reverse (5'-gcatgttgtactttctgctcca-3') primers in exon1 and exon2, respectively.
To check the expression of p53 in morphants by qPCR, primers were as follows: forward: 5'-cgcagccgctatgagatgatt-3'; reverse: 5'-cacttgcggcacttaatggt-3'. The relative expression of p53 was normalized to Histone4 expression (H4) for which primers were as follows: forward: 5'-ggtgatgccctggatgttgt-3'; reverse: 5'-ggcaaaggaggaaaaggactg-3'.
Immunostainining on Xenopus embryos
Embryos were fixed in 4% paraformaldehyde (PFA) overnight at 4°C and stored in 100% methanol at −20°C. Embryos were rehydrated in PBT and washed in MABX (Maleic Acid Buffer + Triton X100 0,1% v/v). Next, embryos were incubated in Blocking Reagent (Roche) 2% BR + 15% Serum + MABX with respective primary and secondary antibodies (see table below). For all experiments secondary antibodies conjugated with Alexa were used. GFPCAAX in Fig. S7E was revealed using a rabbit anti-GFP antibody together with a secondary antibody coupled to HRP, which was revealed as described previously 10. To mark cortical actin in MCCs, embryos were fixed in 4% paraformaldehyde (PFA) in PBT (PBS + 0,1% Tween v/v) for 1h at room temperature (RT), washed 3x10 min in PBT at RT, then stained with phalloidin-Alexa Fluor 555 (Invitrogen, 1:40 in PBT) for 4 h at RT, and washed 3x10 min in PBT at RT. Primary antibodies: mouse anti-Acetylated–α-Tubulin (Clone 6-11B-1, Sigma-Aldrich, 1:1000), rabbit anti-η-Tubulin (Abcam, 1:500), mouse anti-η-Tubulin (Clone GTU88, Abcam, 1:500), Chicken anti-GFP (2B scientific, 1:1000), rabbit anti-GFP (Torrey Pines Biolabs, 1:500). Secondary antibodies: Alexa Fluor 647 goat anti-mouse IgG2a (1:500; Thermo Fisher Scientific), Alexa Fluor 488 goat anti-chicken (1:500; Thermo Fisher Scientific), Alexa Fluor 568 goat anti-rabbit (1:500; Thermo Fisher Scientific).
In situ hybridization on Xenopus embryos
Whole-mount chromogenic in situ hybridization was performed as described previously 43. Whole-mount fluorescent in situ hybridisation (FISH) was performed as described previously 44. For single staining, all RNA probes were labeled with digoxigenin. For FISH on section, embryos were fixed in 4% paraformaldehyde (PFA), stored in methanol for at least 4 h at −20°C, then rehydrated in PBT (PBS + Tween 0.1% v/v), treated with triethanolamine and acetic anhydride, incubated in increasing sucrose concentrations and finally embedded with OCT (VWR Chemicals). 12µm-thick cryosections were made. Double FISH on sections was an adaptation of the whole-mount FISH method. 80ng of cdc20b digoxigenin-labeled sense and antisense riboprobes and 40ng of antisense α-tubulin fluorescein-labeled riboprobe 45 were used for hybridization. All probes were generated from linearized plasmids using RNA-labeling mix (Roche). FISH was carried out using Tyramide Signal Amplification – TSA TM Plus Cyanine 3/Fluorescein System (PerkinElmer). Antibodies: Anti-rabbit-HRP (Interchim, 1:5000), Anti-DigAP (Roche, 1:5000), Anti-DigPOD (Roche, 1:500), Anti-FluoPOD (Roche, 1:500).
Microscopy
Confocal: Flat-mounted epidermal explants were examined with a Zeiss LSM 780 confocal microscope. Four-colors confocal z-series images were acquired using sequential laser excitation, converted into single plane projection and analyzed using ImageJ software. Scanning Electron Microscopy (SEM): skin epidermis of Xenopus embryos from stage 37 was observed and analyzed into a digital imaging microscope (FEI TENEO). Embryos were processed as described previously 44. Transmission Electron Microscopy (TEM): St25 embryos were fixed overnight at 4°C in 2.5% glutaraldehyde, 2% paraformaldehyde, 0.1% tannic acid in a sodium cacodylate buffer 0.05 M pH7.3. Next, embryos were washed 3x15 min in cacodylate 0.05 M at 4°C. Post-fixation was done in 1% osmium buffer for 2 h. Next, embryos were washed in buffer for 15 min. Then, embryos were washed in water and dehydrated conventionally with alcohol, followed by a step in 70% alcohol containing 2% uranyl during 1 to 2 h at RT, or overnight at 4°C. After 3 times in 100% alcohol, completed with 3 washes of acetone. Next, embryos were included in classical epon resin, which was polymerized in oven at 60°C for 48 h. Sections of 80 nm were made and analyzed into an FMI TECNAI microscope with acceleration of 200kV.
Statistical analysis of Xenopus experiments
To quantify the effect of our different experiments, we applied One-way ANOVA analysis and Bonferroni’s multiple comparisons test (t test). ***P<0.05; ns = not significant. Statistical analyses were done using GraphPad Prism 6.
Fig. 3J: 10 cells per condition were analyzed and the total number of η-tubulin positive spots per injected cell was counted.
Fig. 4N: 5 fields (20x zoom) per condition were analyzed, and the total number of properly ciliated MCCs based on acetylated α-tubulin staining among GFP positive cells per field was counted. Each field corresponded to a different embryo.
Acknowledgements
We are grateful to Chris Kintner, Marc Kirschner, Olaf Stemmann, Reinhard Köster, Xavier Morin and Xueliang Zhu for reagents. Imaging in IBDM was performed on PiCSL-FBI core facility, supported by the French National Research Agency through the program "Investments for the Future" (France-BioImaging, ANR-10-INBS-04). Sequencing at UCAGenomiX (IPMC), a partner of the National Infrastructure France Génomique, was supported by Commissariat aux Grands Investissements (ANR-10-INBS-09-03, ANR-10-INBS-09-02) and Canceropôle PACA. The authors thank Florian Roguet for Xenopus care, and Nathalie Garin from Leica Microsystems GmbH for technical advice on STED microscopy. We are grateful to Rainer Waldmann, Kévin Lebrigand and Nicolas Nottet for fruitful discussions on single cell RNA sequencing. We thank Julien Royet and Harold Cremer for insightful comments on the manuscript. This project was funded by grants from ANR (ANR-11-BSV2-021-02, ANR-13-BSV4-0013, ANR-15-CE13-0003), FRM (DEQ20141231765, DEQ20130326464), Fondation ARC (PJA 20161204865, PJA 20161204542), the labex Signalife (ANR-11-LABX-0028-01), and the association Vaincre la Mucoviscidose (RF20140501158, RF20120600738, RF20150501288). OM, CB and DRR were supported by fellowships from Ligue Nationale contre le Cancer (OM and CB), and Fondation ARC (DRR).