RFX2 is broadly required for ciliogenesis during vertebrate development

Abstract

In Caenorhabditis elegans, the RFX (Daf19) transcription factor is a major regulator of ciliogenesis, controlling the expression of the many essential genes required for making cilia. In vertebrates, however, seven RFX genes have been identified. Bioinformatic analysis suggests that Rfx2 is among the closest homologues of Daf19. We therefore hypothesize that Rfx2 broadly controls ciliogenesis during vertebrate development. Indeed, here we show that Rfx2 in Xenopus is expressed preferentially in ciliated tissues, including neural tube, gastrocoel roof plate, epidermal multi-ciliated cells, otic vesicles, and kidneys. Knockdown of Rfx2 results in cilia-defective embryonic phenotypes and fewer or truncated cilia are observed in Rfx2 morphants. These results indicate that Rfx2 is broadly required for ciliogenesis in vertebrates. Furthermore, we show that Rfx2 is essential for expression of several ciliogenic genes, including TTC25, which we show here is required for ciliogenesis, HH signaling, and left-right patterning.

Fig. 1

Rfx2 is expressed in ciliated tissues. (A) Early expression of Rfx2 in epidermis. (B) Rfx2 is expressed in the neural plate. Dorsal view. (B′) (C) Sagittal sections showed that Rfx2 is expressed in the gastrocoel roof plate (GRP) (arrows). BP: blastopore. (D) Transverse section view revealed robust Rfx2 expression in the neural tube. (E) Rfx2 is expressed in the neural tube. Dorsal view. (F) Rfx2 is expressed in otic vesicles and kidneys. (G) Punctate expression of Rfx2 in the epidermis. (H) Closer view of epidermal Rfx2 expression in (G). (I) Epidermal ciliated cells are observed by staining with α-tubulin. Ciliated cells are co-localized with Rfx2 in situ pattern in (H).

Fig. 2

Rfx2 is required for neural tube closure and proper neural cilia formation. (A–D) Stage 19 embryos, dorsal view, anterior top. Black dashed lines outline the distance between the neural folds. (A) Stage 19 control embryo. The neural tube is almost closed. (B) Embryo injected with Rfx2 morpholino dorsally. The neural tube closure defects are shown. The neural tube closure defects caused by disruption of Rfx2 can be partially rescued by co-injection with 150 pg GFP-Rfx2 mRNA (C) and 300 pg GFP-RFX2 mRNA (D). The average distance between the neural folds is shown in (E). (F) Transverse section view of neural plate of stage 18 control embryos. Cilia are stained with acetylated α-tubulin. (G) Transverse section view of neural plate of Rfx2 morphants. Shorter cilia are observed. (H) Cilia were measured in 15 µm projection confocal images. The average length of control cilia is 1.19 ± 0.021 µm (mean ± SEM, n = 225). However, the average length significantly reduces to 0.79 ± 0.016 µm in Rfx2 morphants (mean ± SEM, n = 195). Horizontal lines indicate the mean, vertical lines SEM, *** p<0.0001 Mann–Whitney test.

Fig. 3

Rfx2 is required for Shh signaling. (A–C) In situ hybridization of a Shh downstream gene, Nkx2.2.Dorsal view, anterior left. Stage 24. Expression of Nkx2.2 in the ventral neural tube (A) is dramatically reduced in the spinal cord of Rfx2 morphants (21/29) (B). The reduced expression of Nkx2.2 can be rescued by co-injection of 300 pg GFP-Rfx2 mRNA (C). (D–F) In situ hybridization of a Shh downstream gene, Vax1. Lateral view, anterior left. Stage 35.Expression of Vax1 in the ventral forebrain (D) is robustly reduced in Rfx2 morphants (16/24) (E) and can be rescued (F).

Fig. 4

Rfx2 is required for GRP cilia assembly and the left–right asymmetry pattern. (A) GRP tissue of a membrane RFP-injected control embryo. Acetylated <α-tubulin labels cilia (green) and RFP labeled cell boundary (red). (B) GRP tissue of a Rfx2 morphant. Membrane-RFP is co-injected as a tracer. (C) GRP tissue of an embryo injected with Rfx2 morpholino and GFP-Rfx2 mRNA. (A′), (B′), (C′) are zoom in view of (A), (B), (C), respectively. (D) Average length of GRP cilia. While the average length of control GRP cilia is 8.59 ± 0.18 µm (mean ± SEM, n = 69), it's significantly reduced to 2.34 ± 0.07 µm in Rfx2 morphants (mean ± SEM, n = 113). The shorter cilia phenotype is partially rescued to 4.74 ± 0.14 µm (mean ± SEM, n = 63) by co-injecting with GFP-Rfx2 mRNA. Horizontal lines indicate the mean, vertical lines SEM, *** p<0.0001 Mann–Whitney test. (E–F, I) Pitx2c expression at stage 26, lateral view. In control embryos, Pitx2c is expressed in the left LPM (black arrow) but not the right LPM (blue arrow). (G–H) In Rfx2 morphants, bilateral LPM of Pitx2c expression is observed. (blue arrow). (I) Quantification of Pitx2c expression patterns in control embryos and Rfx2 morphants.

Fig. 5

Rfx2 is essential for epidermal ciliogenesis. (A) A stage 27 control embryo injected with membrane-RFP. Acetylated α-tubulin labels cilia (green) and RFP labels cell boundary (red). (A′) Zoom in view of (A). (B) A stage 27 Rfx2 morphant. (B′) Zoom in view of (B). Note that only a few short axonemes are shown in the Rfx morphant.

Fig. 6

Rfx2 is required for the expression of a ciliary gene TTC25, but not α-tubulin. (A)–(D) Rfx2 morpholino is injected into one ventral blastomere at the 4 cell stage. (A) α-tubulin in situ hybridization on the control side of epidermis. (B) α-tubulin in situ hybridization on the Rfx2 morpholino-injected side. α-tubulin expression is not significantly changed (also see suppl. Fig. 4). (C) TTC25 expression patterns on control epidermis. (D) TTC25 in situ hybridization on Rfx2 morpholino-injected side of epidermis. TTC25 expression is reduced in epidermal multi-ciliated cells. (E)–(H) Rfx2 morpholino is injected into both dorsal blastomeres at the 4 cell stage. (E) Dorsal view of TTC25 expression on neural plate of a control embryo. (F) Dorsal view of TTC25 expression on neural plate of an Rfx2 morphant. TTC25 expression is reduced. (G) Sagittal section view of a control embryo. TTC25 is expressed in the gastrocoel roof plate (GRP). (H) Sagittal section view of a Rfx2 morphant. TTC25 expression is reduced in GRP.

Fig. 7

TTC25 is required for proper neural tube cilia formation and Shh signaling pathway. (A–B) Nkx2.2 expression at stage 24, dorsal view, anterior left. Expression of Nkx2.2 in the ventral neural tube (A) is reduced in the spinal cord of TTC25 morphants (B). (C–D) Vax1b expression at stage 35, lateral view, anterior left. Expression of Vax1b in the ventral fore-brain (C) is reduced, but still present, in TTC25 morphants (D). (E–F) Confocal Z-projections of stage 24 neural tube sections stained for cilia markers acetylated α-tubulin (green) and Arl13b (red). (E) In control embryos, most cilia project into the lumen of the neural tube; ventral neural cilia are longer than dorsal ones and express less Arl13b. (F) In TTC25 morphants, the neural tube fails to close and cilia are not found at the luminal boundary (top). Cilia are fewer, shorter, and stain more intensely with Arl13b. (G) Ventral cilia were measured in 3 sections (as in E, F) from 5 embryos each. The 20 longest ventral cilia from each embryo are plotted. Horizontal lines indicate the mean, vertical lines SEM, p<0.0001 Mann–Whitney test.

Fig. 8

Rfx2 is required for GRP cilia assembly and the left–right asymmetry. (A–B) Pitx2c expression at stage 26, lateral view. In wildtype embryos, Pitx2c is expressed in the left LPM (A, A′, C). In TTC25 morphants, only 33% of embryos show normal left LPM expression (C), and Pitx2c is often expressed in the right LPM (B′, C). (D–E) Confocal images of the posterior GRP (node) stained with acetylated α-tubulin to visualize cilia and injected with membrane-RFP to outline cells. (F) GRP cilia were measured in confocal images as in D, E for 5 embryos each. Horizontal lines indicate the mean, vertical lines SEM, p<0.0001 Mann–Whitney test.