Systematic discovery of novel ciliary genes through functional genomics in the zebrafish

Cilia are microtubule-based hair-like organelles that play many important roles in development and physiology, and are implicated in a rapidly expanding spectrum of human diseases, collectively termed ciliopathies. Primary ciliary dyskinesia (PCD), one of the most prevalent of ciliopathies, arises from abnormalities in the differentiation or motility of the motile cilia. Despite their biomedical importance, a methodical functional screen for ciliary genes has not been carried out in any vertebrate at the organismal level. We sought to systematically discover novel motile cilia genes by identifying the genes induced by Foxj1, a winged-helix transcription factor that has an evolutionarily conserved role as the master regulator of motile cilia biogenesis. Unexpectedly, we find that the majority of the Foxj1-induced genes have not been associated with cilia before. To characterize these novel putative ciliary genes, we subjected 50 randomly selected candidates to a systematic functional phenotypic screen in zebrafish embryos. Remarkably, we find that over 60% are required for ciliary differentiation or function, whereas 30% of the proteins encoded by these genes localize to motile cilia. We also show that these genes regulate the proper differentiation and beating of motile cilia. This collection of Foxj1-induced genes will be invaluable for furthering our understanding of ciliary biology, and in the identification of new mutations underlying ciliary disorders in humans.

were designed against 48 of the 50 randomly selected FIGs. Each morpholino was injected into zebrafish embryos and RNA was extracted at 24 hpf in order to verify that the morpholino is able to cause mis-splicing. PCR was used to check the resultant cDNA near the predicted morpholino binding site. Comparing the wild type (wt) amplification to the morphant (mo), we found that all of the morpholinos are able to induce either a substantial amount of mis-splicing or the degradation of the endogenous transcript. Reactions without reverse transcriptase are included (no RT) to eliminate the possibility of genomic DNA contamination, while actb1 primers were used as loading controls (LC) to ensure equal amounts of starting material.
Bands marked with a red asterisk were cloned and sequenced in order to verify the inclusion/exclusion of intronic/exonic sequence. The band marked with a white asterisk in tmc5 morphants showed similar levels to wild type embryos; however, these embryos also displayed a prominent mis-spliced product.

Supplementary Movie 7. KV motile cilia in a lace1b morpholino-injected embryo expressing Arl13b-GFP
A sample movie of KV cilia motility in a lace1b morpholino-injected embryo at 14-15 hpf with GFP-labelled cilia. Video acquisition was at 321 frames/s and playback is at 11.4 frames/s.

Supplementary Movie 8. KV motile cilia in a tp53bp2 morpholino-injected embryo expressing Arl13b-GFP
A sample movie of KV cilia motility in a tp53bp2 morpholino-injected embryo at 14-15 hpf with GFP-labelled cilia. Video acquisition was at 334 frames/s and playback is at 11.8 frames/s.

Supplementary Movie 9. KV motile cilia in a kcnip1a morpholino-injected embryo expressing Arl13b-GFP
A sample movie of KV cilia motility in a kcnip1a morpholino-injected embryo at 14-15 hpf with GFP-labelled cilia. Video acquisition was at 159 frames/s and playback is at 5.6 frames/s.

Supplementary Movie 10. KV motile cilia in an aftphb morpholino-injected embryo expressing Arl13b-GFP
Development | Supplementary Material 9 A sample movie of KV cilia motility in an aftphb morpholino-injected embryo at 14-15 hpf with GFP-labelled cilia. Video acquisition was at 334 frames/s and playback is at 11.8 frames/s.

Supplementary Movie 11. KV motile cilia in a tmc5 morpholino-injected embryo expressing Arl13b-GFP
A sample movie of KV cilia motility in a tmc5 morpholino-injected embryo at 14-15 hpf with GFP-labelled cilia. Video acquisition was at 378 frames/s and playback is at 13.4 frames/s.

Supplementary Movie 12. KV motile cilia in a spa17 morpholino-injected embryo expressing Arl13b-GFP
A sample movie of KV cilia motility in a spa17 morpholino-injected embryo at 14-15 hpf with GFP-labelled cilia. Video acquisition was at 484 frames/s and playback is at 17.1 frames/s.

Supplementary Movie 13. KV motile cilia in an arhgef18b morpholino-injected embryo expressing Arl13b-GFP
A sample movie of KV cilia motility in an arhgef18b morpholino-injected embryo at 14-15 hpf with GFP-labelled cilia. Video acquisition was at 156 frames/s and playback is at 5.5 frames/s.

Supplementary Movie 14. KV motile cilia in an ak9 morpholino-injected embryo expressing Arl13b-GFP
A sample movie of KV cilia motility in an ak9 morpholino-injected embryo at 14-15 hpf with GFP-labelled cilia. Video acquisition was at 139 frames/s and playback is at 4.9 frames/s. Table S1 Symbol and version. This is followed by the verterbrate ortholog Ensembl Gene ID, Symbol and Description. An internal CG code is also given (CG0XX) for reference.

Supplementary
Sheet 1 (Zebrafish qPCR) includes the zebrafish qPCR forward primer name and sequence, followed by the qPCR reverse primer name and sequence for each gene.
The last three columns of this sheet contain the fold change increase in Foxj1 overexpressed embryos relative to wild type, the standard error of the mean (SEM) for each fold change, and the p-value when a t-test is performed comparing the gene Development | Supplementary Material expression levels in the Foxj1 overexpressed embryos to the levels in wild type. All of the data were normalized against expression levels of a housekeeping gene, rplp0, and actb1 levels were checked as a negative control (highlighted in green). Sheet 2 (RACE and Localization Results) lists the results of the RACE and sequence analysis along with the protein localization results. Controls include embryos injected with mRNA encoding GFP alone and immunofluorescence analysis of wild type, uninjected embryos (highlighted in green). Sheet 3 (Morpholino Information) lists the morpholino version (an internal label), the sequence of each antisense morpholino (5' to 3'), the transcript against which the morpholino was designed, and the exon/intron junction of the canonical transcript which the morpholino targets.
The next two columns contain the primer names and sequences for the oligonucleotides used to test the mis-splicing for each morpholino (forward and reverse primers, respectively). Finally, the amount of morpholino (in picomoles) injected into each embryo for phenotypic analysis is given along with the results of the PCR and sequencing. Sheet 4 (Morpholino Phenotypes) lists the counts for each of the phenotypes assayed during the morpholino analysis. The numbers of affected, normal, and total embryos are given for each phenotype, in addition to the percentage of affected embryos and the p-value of the Fisher's exact test comparing the morpholino-injected embryos with wild type embryos. Phenotypes analyzed include: curved body axis (CBA), otolith counts, hydrocephalus, kidney cysts and left-right asymmetry. The standard control morpholino-injected and wild type embryo analyses are highlighted in green. Sheet 5 (Cilia Length and Motility) Development | Supplementary Material 13 includes average length measurement of pronephric cilia along with standard deviations and p-values. In addition, KV cilia motility data is provided for the ten assayed morpholinos, including average percent motility and average beat frequency, standard deviations and p-values. Control morpholino-injected embryos and wild type embryos are reported as controls (green). Sheet 6 (Human qPCR) includes the human qPCR forward primer name and sequence, followed by the qPCR reverse primer name and sequence for each gene. The next column contains the results of the expression analysis by qPCR for human tissues. The final three columns contain the fold change increase in human brain, lung, testis or heart relative to skeletal muscle, the standard error of the mean (SEM) for each fold change and the p-value when a t-test was performed comparing the gene expression levels in the tissue in question relative to muscle. All of the data were normalized against expression levels of a housekeeping gene, ACTB, while GAPDH levels were