HEATR2 plays a conserved role in assembly of the ciliary motile apparatus

Abstract

Cilia are highly conserved microtubule-based structures that perform a variety of sensory and motility functions during development and adult homeostasis. In humans, defects specifically affecting motile cilia lead to chronic airway infections, infertility and laterality defects in the genetically heterogeneous disorder Primary Ciliary Dyskinesia (PCD). Using the comparatively simple Drosophila system, in which mechanosensory neurons possess modified motile cilia, we employed a recently elucidated cilia transcriptional RFX-FOX code to identify novel PCD candidate genes. Here, we report characterization of CG31320/HEATR2, which plays a conserved critical role in forming the axonemal dynein arms required for ciliary motility in both flies and humans. Inner and outer arm dyneins are absent from axonemes of CG31320 mutant flies and from PCD individuals with a novel splice-acceptor HEATR2 mutation. Functional conservation of closely arranged RFX-FOX binding sites upstream of HEATR2 orthologues may drive higher cytoplasmic expression of HEATR2 during early motile ciliogenesis. Immunoprecipitation reveals HEATR2 interacts with DNAI2, but not HSP70 or HSP90, distinguishing it from the client/chaperone functions described for other cytoplasmic proteins required for dynein arm assembly such as DNAAF1-4. These data implicate CG31320/HEATR2 in a growing intracellular pre-assembly and transport network that is necessary to deliver functional dynein machinery to the ciliary compartment for integration into the motile axoneme.

Figure 1. HEATR2 splice mutation results in alteration of the final conserved HEAT repeat and protein instability.

(A) Pedigree of related families of UK-Pakistani descent. IV∶4 identifies the proband, also designated by the arrow. Solid symbols (individuals IV∶1, IV∶4 and IV∶10) indicate those affected with PCD. Double lines indicate consanguineous marriages. The individuals labeled DNA signified those that had their DNA included in SNP genotyping. (B) Schematic of HEATR2 transcript showing the transversion mutation (ENST00000297440:c.2432-1G>C) affecting the splice acceptor site of the final exon. The mutation results in inactivation of this splice site and utilization of an adjacent downstream cryptic splice acceptor site in exon 13, causing a 2-nucleotide AG deletion in the HEATR2 transcript, resulting in a frameshift in translation (Figure S2B). This is predicted to alter the final 44 amino acids of the protein and add an additional 33 amino acids with creation of a novel termination signal at codon 888 in the 3′UTR (See Figure S3A). This mutation disrupts the final highly conserved HEAT repeat and alters the C-terminus of the ARM-type fold superfamily domain (red). (C) Relative expression levels of HEATR2 transcript by RT-qPCR, when normalized to the reference TBP gene. (D) The PCD transversion mutation (ENST00000297440:c.2432-1G>C) does not affect HEATR2 transcript stability or gross splicing as shown by RT-PCR on parental control (C) and patient (P*) cDNA from LCLs. PCR products spanning the gene including the splice acceptor mutation at Exon 11–13 and Exon 12-3′UTR show no obvious alterations in size. Direct sequencing confirmed a 2 base pair deletion consistent with efficient splicing to the cryptic splice acceptor at the start of exon 13 in PCD patients (Figure S2B). (E) Western blot analysis on total protein extracts from unrelated control, heterozygous parental and homozygous patient LCLs demonstrates the PCD mutation (ENST00000297440:c.2432-1G>C) results in an elongated HEATR2 protein present at reduced levels implying instability. The slight shift in mobility of the protein in the patient is consistent with the predicted 3 kDa size shift due to the amino acid alterations described. β-actin is used as a loading control. (For longer exposure see Figure S3B). (F) Levels of HEATR2 protein normalized relative to β-actin reveal that parental samples which are heterozygous for the mutation shows a reduction to ≈50% of that of unrelated controls whilst the homozygous patient sample shows a reduction to ≈3% of control levels.

Figure 2. The HEATR2 orthologue CG31320 is highly expressed in motile ciliated mechanosensory Ch neurons in an Rfx- and Fox-dependent manner.

(A–D) In situ hybridization shows that CG31320 mRNA is present in chordotonal (Ch) neurons from about stage 12, when transient mesoderm expression is also observed (A). Through early neuronal differentiation (stage 14: B, B′ (higher magnification) to late neuronal differentiation (stage 16: C, C′), CG31320 expression is highly expressed in a restricted pattern to Ch neurons. (C″: higher magnification of C, stage 16). Overlay of a Ch neuron schematic illustrating strong and restricted CG31320 expression in these clusters of ciliated mechanosensory neurons in late stage embryos. (Scale bars: B, B′,C: 100 µm; C′,C″: 20 µm) (D) Double labeling of stage 16 wild type embryos for anti-Rfx (red, nuclear stain) and CG31320 mRNA (diffuse blue) shows co-labelling between Rfx and CG31320 in late Ch neurons. The master ciliogenic transcription factor Rfx is expressed transiently in all sensory neurons but at this stage it is only present in the motile ciliated Ch neurons. (Scale bar: 100 µm). (E,F) CG31320 expression is dependent on the Rfx transcription factor. CG31320 mRNA expression (blue) is lost in Ch neurons (anti-HRP: neuronal glycoproteins, red) of rfx⁴⁹ mutants . (Scale bar: 20 µm). (G–I) Fd3F is necessary and sufficient to drive CG31320 expression in sensory neurons. Compared to control (G), CG31320 expression is abolished in homozygous fd3F¹ mutant embryo (H). (I) Conversely, ectopic fd3F expression (scaGal4, UAS-fd3F) expands CG31320 expression into other ciliated sensory neurons in stage 16 embryos. (Scale bar: 100 µm).

Figure 3. CG31320 is required for mechanosensory structure and function in Ch neurons.

(A–C) Control and (D–F) Sensory-neural specific CG31320 RNAi knock-down (UAS-Dcr2; scaGal4/UAS-CG31320 RNAi). (A,D) In-situ hybridisation confirms that substantial loss of CG31320 mRNA was achieved by the knock-down. (Scale bar: 50 µm). (B,E) Immunofluorescence of larval Ch neurons using the pan-neuronal marker (anti-HRP: green, neuronal glycoproteins marking luminal bands at level of basal body and close to ciliary dilation) and the ion channel NompC/TRPN1, (anti-NompC; magenta; marks the distal non-motile cilium tip) shows that loss of CG31320 results in no gross cilia dysmorphology or loss of compartmentalization of ciliated Ch structures. (Scale bar:10 µm). (C,F) TEM of Ch cilia cross-sections from adult antennae (Johnston's organ), showing nine axonemal microtubule doublets shown schematically in (G). (C) The electron-dense structures corresponding to inner (red arrowhead) and outer (blue arrowhead) axonemal dynein arms are clearly seen in wild-type. (F) These are not observed in CG31320 knock-down cilia. (Scale bar: 100 nm). (G) Schematic illustration of Drosophila Ch neurons showing the localisation of markers in the cilia and the presence of dynein arms in the proximal motile zone. (H) Ch neuronal function is measured by the negative geotaxis climbing assay for adult flies. The height climbed by control (n = 43) versus CG31320 RNAi flies (n = 64) reveals the latter to be uncoordinated. (Mann-Whitney U test: P≤0.0001). (I) CG31320 RNAi knock-down larvae do not respond in an auditory assay. Retraction score is the number of larvae (in a sample of 5) exhibiting head shortening during a 0.5 second time window. Shown is the mean retraction score for several tests (control: n = 15; RNAi: n = 16); error bars are standard error of the mean. Statistical analysis was performed using the Friedman test followed by Dunn's multiple comparison post-hoc test. (** represents P≤0.01; **** represents P≤0.0001).

Figure 4. CG31320 is required for sperm flagellar motility and male fertility.

(A,C,E) Control and (B,D,F) CG31320 testes-specific inducible RNAi knock-down (UAS-Dcr2; UAS-CG31320 RNAi; Bam-VP16-Gal4) adult flies. (A,B) Upon knock-down, gross morphology of the testis and seminal vesicle (SV) is normal and sperm bundles can be observed in control and knock-down testes (arrows). (C,D) Higher magnification views show normal organization of developing sperm bundles in the proximal testis in CG31320 knock-down. (E,F) In control (E), the seminal vesicle is full of mature sperm and many motile spermatozoa are visible swimming away, (F) Mature sperm are not visible within CG31320 knock-down seminal vesicles and no motile sperm are observed upon its dissection. (G–I) TEM images of adult testes post-elongation flagellar transverse sections show (G) a control spermatid, with dynein arms visible on some of the microtubule doublets (colored arrowheads) whilst (H) CG31320 RNAi mutants lack dynein arms. Despite a normal “9+2” configuration, some mutant “A” doublet sub-tubules have electron-dense cores (white arrowheads) or disruptions suggesting defects in nexin links between AB doublets (arrows). (I) Transverse section of CG31320 RNAi knock-down mutant spermatid cyst highlights frequency of axonemal disruption (arrows). Scale bars: 100 nm (G,H), 2 µm (I). (J) Summary of male fertility phenotypes between control and CG31320 testes-specific RNAi knock-down. Number of males producing progeny represents progeny from crosses to wild-type females. The knock-down flies were completely infertile, even though mating was observed. In addition to empty seminal vesicles, knock-down testes exhibit accumulation of sperm and some disruption of sperm bundles. These phenotypes appear to represent secondary consequences of the failure of sperm to move to the seminal vesicle, which requires sperm motility.

Figure 5. CG31320/HEATR2 is conserved in eukaryotes with motile cilia/flagella and its associated axonemal dynein apparatus.

CG31320/HEATR2 orthologues are found in species with cilia/flagella that have motile function and retain elements of the axonemal dyneins required for this motility. Species which have no cilia (ie. amoebozoans, flowering plants, yeast) or those which lack motile cilia (i.e. nematodes) have lost HEATR2 orthologues as well as the axonemal dynein genes. Interestingly, unusual species with variant motility programmes still retain HEATR2 orthologues. These include T. pseudonana whose male gametes have motile axonemes without inner arm dyneins, and P. patens, whose male gametes have motile flagella without outer arm dyneins. Similarly, P. falciparum which assembles its flagella intracytosolically through an IFT-independent programme, retains a HEATR2 orthologue. This suggests CG31320/HEATR2 is an essential element of an ancient programme required for ciliary/flagellar motility. This figure is a summary of a more extensive search detailed in Table S1 for CG31320/HEATR2 orthologues as well as axonemal dynein components of the outer (ODA) and inner (IDA) dynein arms, as summarized in columns. Filled circles: orthologues as determined by the top score in reciprocal BLASTP or TBLASTN searches. Open circle: no homologue present. Half-filled circle: evidence supporting existence of at least one orthologue per category as analyzed in Table S1 by reciprocal BLASTP or TBLASTN searches. Information for the intraflagellar transport (IFT) pathway built upon Wickstead and Gull (2007) with our own searches for IFT components by reciprocal BLASTP or TBLASTN searches.

Figure 6. CG31320/HEATR2 orthologues share conserved upstream regulatory FOX motifs and X-boxes of a master cilia motility transcriptional programme.

(A) Using the human upstream epigenetic markings and conservation to mouse and rat to define conserved predicted regulatory elements, we focused analysis on the 500 bp upstream of the HEATR2 ATG and syntenic regions in other species to identify X-box sequences, along with the nearest conserved FOX motifs. These sequences are coloured where they conform to recognized core consensus sequences for generic FOX proteins (RYMAAYA [71]) and RFX (RYYRYYN_(1–3)RRNRAC [42]). Nucleotides are shown in grey if they vary from the consensus. Note for the second identified X-box site the 3′ site is extremely well-matched whilst the 5′ half-site is often more degenerate , . The distance from the Fox motif and X-box to the transcription start site is indicated, or else the distance to the ATG is indicated if a sizeable 5′UTR is present (i.e. D. melanogaster, C. lupus). An expanded table of the analysis is provided in Table S2. (B) ChIP-Seq data reveals a single, specific RFX3 peak 200 bp upstream from the transcriptional start site in OF1 mouse primary differentiated ependymal cell culture. Insert illustrates the two X-boxes bioinformatically predicted within the peak sequence. (C) Directed ChIP-qPCR data validates RFX3 occupancy is enriched at Heatr2 promoter in OF1 cells, normalized to known target gene Dyn2li1 and relative to a control sequence, downstream region in the Dync2li1 gene. (D) Heatr2 expression is ≈55% reduced in Rfx3^−/− ependymal cells similar to reductions in expression observed for two known direct Rfx3 targets, Dync2li1 and Bbs5. qPCR data represent the average of three different assays performed in triplicate ± SEM. All data are considered significant using Student's t-test. (Heatr2 P = 0.003652632; Dync2li P = 0.013123897; Bbs5 P = 0.022511438).

Figure 7. HEATR2 is highly expressed in tissues with motile cilia.

(A) Developmental changes in gene expression were assayed by RT-qPCR on RNA extracted from wild-type mouse lungs and trachea from E14.5-P2 (N = 3 independent biological samples for each time-point). qPCR data represents the average of three different assays for three samples performed in triplicate ± SEM. Kruskal-Wallis non-parametric analysis of variance was performed and was significant for all genes (Zmynd10 P = 1.558e⁻⁰⁸; Dnahc5 P = 1.141e⁻⁰⁵; Dnali1 P = 9.814e⁻⁰⁶; Foxj1 P = 1.956e⁻⁰⁹; Heatr2 P = 1.604e⁻⁰⁶; Rfx3 P = 6.68e⁻⁰⁶). (B,C) Immunostaining on sections of E15.5 mouse lungs, where strong RFX3 and FOXJ1 signals are co-expressed in a “salt and pepper” pattern only in large proximal airways, not smaller, more distal airways (see Figure S6A,B). Although they are not yet multiciliated, these cells also express components of axonemal dyneins and high levels of HEATR2 in their cytoplasm. (Scale bar: B,C = 50 µm, B′,C′ = 10 µm). (D–G) Immunostaining of human nasal brush epithelial cells for: (D) RFX3 (HPA: red), acetylated tubulin (green) and DNALI1 (SC: purple); (E) HEATR2 (Novus: red), FOXJ1 (green) and DNALI1 (SC: purple); and (F,G) HEATR2 (Proteintech: red), FOXJ1 (green) and RFX3 (SC: purple). White arrowheads highlight fully mature motile, multiciliated cells (MMCs) that express lower nuclear RFX3 and FOXJ1 with reduced HEATR2 and with axonemal dynein components entirely in cilia. Arrows highlight immature MMCs for comparison. HEATR2 is entirely cytoplasmic at all stages examined (Scale bar: D–G = 10 µm).

Figure 8. Cytoplasmic HEATR2 is expressed during early ciliogenesis.

(A–C) Immunostaining of control human nasal brush epithelia reveals endogenous HEATR2 (red: Novus (A), Proteintech (C)) is highly enriched in the cytoplasm of developing MMC when components of outer dynein arms (B: DNAH5, red; C: DNAI2: green) as well as inner dynein arms (A–C: DNALI1, purple) are predominantly cytoplasmic. Arrowheads highlight fully mature MMCs where these components are exclusively axonemal and with relatively lower levels of HEATR2. Arrows highlight immature MMCs for comparison. Nuclei are stained with DAPI (blue). (Scale bar: A–C, 10 µm) (D) Double immunofluorescence of 22C10 (magenta: Futsch, cytoplasmic/membrane marker, but not cilium, of all sensory neurons) and CG31320::mVenus (green) indicates there is cytoplasmic but no ciliary localization of CG31320 in stage 16 Ch neurons (Ci: cilia, marked with square bracket). As this construct uses the upstream regulatory region of CG31320 containing the X and Fox motifs to drive reporter expression, it further supports regulation occurs via these sites. (See Figure S7).

Figure 9. Cytoplasmic HEATR2 is required for the pre-assembly of axonemal dynein machinery necessary for motility.

(A) Immunofluorescence for axonemal dynein heavy chain 5 (DNAH5: green) on respiratory cells from patients with HEATR2 mutations compared to non-related control cells, shows loss of type 1 and type 2 DNAH5-positive staining although axonemes are still present (acetylated tubulin:red). Nuclei are stained with DAPI. (Scale bar: 10 µm). (B) Extracts prepared from control human terminally differentiated respiratory airway cultures (40 days ALI, Epithelyx) were subjected to immunoprecipitation (IP) with antibodies to HEATR2 (Proteintech) or control rabbit immunoglobulin G (GFP). Resulting immunocomplexes (IP: right) as well as dilutions of original extract (INPUT: left) were subjected to immunoblot analysis with antibodies to HEATR2 (Proteintech) or DNAI2 (Abnova). (See also Figure S8). (C) No staining of axonemal dynein light intermediate chain 1 (DNALI1: green) is observed in patients with HEATR2 mutations. No signal above background is detected in patient cells, in contrast to strong axonemal localization in non-related control cells. Nuclei are stained with DAPI. (Scale bar: 10 µm). (D) The DNALI1 orthologue in fly, CG6971::mVenus (green), fails to localize to ciliary axonemes (Ci: cilia, marked with square bracket) of Ch neurons (magenta: 22C10/Futsch) in CG31320 knock-down larvae. (E) Schematic of dynamic role of HEATR2 in developing airway epithelial MMCs. Progenitor cells exit the cell cycle to commit to the MMC lineage with primary cilia. These cells express low levels of RFX3 (light red nuclei). Other upstream factors governing multiciliogenesis (MCN, MYB) induce centriole amplification as well as expression of FOXJ1 (bright green nuclei), required for centriole docking. High FOXJ1 and RFX3 drive a cilia motility transcriptional cascade leading to high expression of HEATR2 as well as expression of axonemal dynein components. HEATR2 is involved in the pre-assembly and/or delivery of future dynein arms to the apical cilia base. In fully mature MMCs, once inner and outer arm dynein complexes are delivered and incorporated into motile ciliary axonemes, relative levels of HEATR2 as well as RFX3 and FOXJ1 are reduced. This conserved regulatory motility module is required to drive high levels of HEATR2 when axonemal dyneins are being assembled and trafficked.