A DEAD-box helicase drives the partitioning of a pro-differentiation NAB protein into nuclear foci

Abstract

How cells regulate gene expression in a precise spatiotemporal manner during organismal development is a fundamental question in biology. Although the role of transcriptional condensates in gene regulation has been established, little is known about the function and regulation of these molecular assemblies in the context of animal development and physiology. Here we show that the evolutionarily conserved DEAD-box helicase DDX-23 controls cell fate in Caenorhabditis elegans by binding to and facilitating the condensation of MAB-10, the C. elegans homolog of mammalian NGFI-A-binding (NAB) protein. MAB-10 is a transcriptional cofactor that functions with the early growth response (EGR) protein LIN-29 to regulate the transcription of genes required for exiting the cell cycle, terminal differentiation, and the larval-to-adult transition. We suggest that DEAD-box helicase proteins function more generally during animal development to control the condensation of NAB proteins important in cell identity and that this mechanism is evolutionarily conserved. In mammals, such a mechanism might underlie terminal cell differentiation and when dysregulated might promote cancerous growth.

Fig. 1. MAB-10 protein forms dynamic nuclear foci.

a Known gene regulatory interactions in the heterochronic pathway. LIN-29 (red) and MAB-10 (blue) act together on a subset of LIN-29 target genes to downregulate larval-specific genes and turn on adult-related genes. b Differential interference contrast (DIC) and confocal GFP fluorescence images of hypodermal cells in early L4, young adult, and 1-day-old adult wild-type animals expressing the translational reporter n5909 [mab-10::gfp], which tags the endogenous mab-10 gene locus with gfp. Images are representative of ten animals. Scale bar, 40 μm. c Yeast two-hybrid spot assay showing the self-interaction of MAB-10 protein (upper right spot) in quadruple drop-out plates (-His/-Ade/-Trp/-Leu) containing the competitive inhibitor of histidine synthesis 3-AT. d Representative confocal fluorescent images of individual frames from in vivo fluorescence recovery after photobleaching (FRAP) time-lapse studies of two MAB-10 nuclear foci (“Focus”) and two diffuse MAB-10::GFP signals (“Diffuse”) in hypodermal nuclei of 1-day-old wild-type adults expressing the reporter n5909 [mab-10::gfp]. Images are representative of n = 13 (for “Focus”) and n = 14 (for “Diffuse”) independent experiments examining the MAB-10::GFP signal. Dotted red circle, nuclear circumference; white circle, focal area of laser bleaching and region of measurement; scale bars, 2 μm. e FRAP curves quantifying observations from the experiments shown in (c) (two MAB-10 nuclear foci (“Focus”, red and pink) and two diffuse MAB-10::GFP signals (“Diffuse”, blue and gray) in hypodermal nuclei). Fluorescence intensity measurements in the area of laser bleaching, normalized by considering the pre-bleach fluorescence intensity to be 1 and post-bleach intensity to be 0, are shown. t = 0, time measuring pre-bleaching intensity. f Half-maximal recovery times from FRAP data for 13 experiments examining MAB-10::GFP in foci (“Focus”) and 14 experiments examining MAB-10::GFP diffuse signal (“Diffuse”). Error bars, mean value +/− SEM. ****P = 1.278 × 10⁻⁶ (two-sided t test). Source data for (e, f) are provided as a Source Data file.

Fig. 2. DDX-23 is evolutionarily conserved and functions in hypodermal stem cell biology.

a Schematic of the F2 non-clonal mab-10 enhancer screen to identify factors that act with or in parallel to MAB-10. mab-10(tm2497) mutants were ethyl methanesulfonate (EMS)-mutagenized, and F2 progeny were screened for second-site mutations that further reduced the low hypodermal col-19 expression levels of mab-10(tm2497) mutants. b The mab-10 enhancer screen isolated two alleles (n5703 and n5705) of ddx-23. The ddx-23(+) transgene nEx2848, which contains wild-type ddx-23, rescued the low maIs105 GFP expression. Representative DIC and GFP fluorescence micrographs of five animals per genotype are shown of C. elegans adults of the genotypes indicated and grown at 20 °C. Scale bar, 500 μm. c Representative DIC and GFP fluorescence micrographs of wild-type and ddx-23 single mutant adult animals grown at 20 °C. Images are representative of five animals per genotype. The ddx-23(+) transgene nEx2958, which contains wild-type ddx-23 tagged with tagRFP-T, rescued the low maIs105 GFP expression. Scale bar, 500 μm. d Schematic of ddx-23 gene structure. DEAD-box domain (orange) and helicase C domain (blue) are depicted. The ddx-23(n5703) and ddx-23(n5705) mutations isolated from the mab-10 enhancer screen are indicated. e Sequence alignments of DDX-23 homologs from S. cerevisiae (Prp28), zebrafish (Ddx23), Mus musculus (DDX23), and Homo sapiens (DDX23). Only the region surrounding amino acid residues G293 and I383, which are altered by the ddx-23 mutations n5703 and n5705, respectively, is shown. Arrows indicate the completely conserved G293 and I383 residues. f Representative micrographs of maIs105[col-19::gfp] reporter GFP expression in 1-day-old ddx-23(n5705) adults with or without the transgene overexpressing human DDX23 (ceDDX23). Images are representative of 5–8 animals per genotype. Scale bar, 500 μm. g Number of seam cells that underwent extra seam cell division, scored in 1-day-old adults in wild-type (n = 8), ddx-23(n5703) (n = 19), ddx-23(n5705) (n = 18), mab-10(tm2497) (n = 36), mab-10(tm2497);ddx-23(n5705) (n = 15), and lin-29(n546) (n = 7) animals. Error bars, mean value +/− SD. ****P < 0.0001 (two-sided t test). P = 1.0 × 10⁻⁵ (wild-type vs. ddx-23(n5703)), P = 1.8 × 10⁻⁶ (wild-type vs. ddx-23(n5705)), P = 7.8 × 10⁻¹⁶ (wild-type vs. mab-10(tm2497)), P = 9.9 × 10⁻¹² (wild-type vs. mab-10(tm2497);ddx-23(n5705)), P = 8.8 × 10⁻⁹ (mab-10(tm2497) vs. mab-10(tm2497);ddx-23(n5705)), and P = 5.6 × 10⁻¹² (ddx-23(n5705) vs. mab-10(tm2497);ddx-23(n5705)). h Proposed role of the ddx-23 gene in the heterochronic pathway. DDX-23 acts with MAB-10 to control seam cell exit from the cell cycle but acts independently of MAB-10 to control seam cell fusion in adults. The human homologs of the C. elegans heterochronic genes are indicated in blue. Source data for (g) is provided as a Source Data file.

Fig. 3. DDX-23 enhances the partitioning of MAB-10 proteins into nuclear foci.

a Representative confocal fluorescent images of 4 animals expressing n5909[MAB-10::GFP] (green channel) and n6092[tagRFP-T::ddx-23] (red channel), which fluorescently tag mab-10 and ddx-23, respectively, at their endogenous loci. Inset: Hypodermal nuclei. Scale bar, 40 μm. b Yeast two-hybrid spot assay showing the interaction of DDX-23 protein fused to the GAL4 activation domain (upper right spot) but not of a control with the GAL4 activation domain-only (“Empty vector”) protein with MAB-10 protein fused to the GAL4 DNA-binding domain in quadruple drop-out plates (-His/-Ade/-Trp/-Leu) containing the competitive inhibitor of histidine synthesis 3-AT. c Schematic representation of the split GFP (spGFP) approach for assessing protein-protein interactions in vivo. Animals expressed an N-terminal GFP fragment (spGFPN) fused to DDX-23 (purple) in combination with a C-terminal GFP fragment (spGFPC) fused to MAB-10 (blue). A physical interaction between the DDX-23::spGFPN and MAB-10::spGFPC proteins leads to the emission of green fluorescence (depicted as a green circle). d Representative confocal images (merged DIC and GFP channels) of spGFP experiments using DDX-23::spGFPN and MAB-10::spGFPC (or spGFPC by itself as a control). Images are representative of 8–10 animals per experimental condition. Scale bar, 20 μm. e Confocal fluorescence images of animals expressing a MAB-10 translational reporter (P_mab-10::mab-10::mCherry::mab-10 3’UTR; red channel), DDX-23::spGFPN and MAB-10::spGFPC (or empty spGFPC as a control; green channel). Image is representative of five animals per condition. Inset: seam cell. Dotted white circle, nuclear circumference; scale bar, 20 μm. f Representative images of in vitro assays testing heterotypic interactions of purified DDX-23-mNeonGreen and MAB-10-mScarlet I (right panel pairs). DDX-23-mNeonGreen can self-interact in the presence of mScarlet I alone (left panel pairs). MAB-10-mScarlet I can self-interact in the presence of mNeonGreen alone (middle panel pairs) but interaction is enhanced in the presence of DDX-23-mNeonGreen (right panel pairs). Images are representative of three independent experiments. Micrographs were color-inverted for better visualization. Scale bar, 200 μm. g Schematic view of control mNeonGreen protein alone, wild-type DDX-23-mNeonGreen, mutant DDX-23(I383F)-mNeonGreen, and a DDX-23 protein lacking its IDR (DDX-23(ΔIDR)-mNeonGreen) used for recombinant protein production. h Representative images of in vitro interaction assays of control mNeonGreen protein alone, wild-type DDX-23-mNeonGreen, mutant DDX-23(I383F)-mNeonGreen, and DDX-23(ΔIDR)-mNeonGreen, each with purified MAB-10-mScarlet I. Images are representative of three independent experiments. Micrographs were color-inverted for better visualization. Scale bar, 50 μm. i Quantification of the number of MAB-10 spots when mixed in heterotypic assays with an mNeonGreen alone control, wild-type DDX-23, mutant DDX-23[I383F] or DDX-23(ΔIDR) in three independent experiments. Error bars, mean value +/− SEM. ***P = 0.0004 (mNeonGreen vs. DDX-23(WT)-mNeonGreen), ***P = 0.0007 (DDX-23(WT)-mNeonGreen vs. DDX-23(ΔIDR)-mNeonGreen) and **P = 0.0024 (DDX-23(WT)-mNeonGreen vs. DDX-23[I383F]-mNeonGreen) (two-sided t tests). j Fraction of animals with discrete MAB-10 foci in 1-day-old wild-type, ddx-23(n5703), and ddx-23(n5705) adults. n = 30 independent animals were scored for each genotype. All strains contain the reporter n5909 [mab-10::gfp]. k DIC and confocal fluorescent images of hypodermal cells in 1-day-old wild-type, ddx-23(n5703) and ddx-23(n5705) adults expressing the reporter n5909 [mab-10::gfp]. Images are representative of 12–15 animals per genotype. Scale bar, 20 μm. l Corrected total fluorescent intensity of MAB-10::GFP in the hypodermal nuclei of 1-day-old wild-type (n = 15) and ddx-23(n5705) (n = 12) adults expressing the reporter n5909 [mab-10::gfp]. ns not significant, P = 0.2381 (two-sided t test). Source data for (i, j, l) are provided as a Source Data file.

Fig. 4. DDX-23-driven formation of MAB-10 (NAB) nuclear foci might control terminal differentiation and the onset of adulthood by transcriptionally repressing LIN-29 (EGR) target hedgehog and larval-specific genes.

a Representative confocal fluorescent images of five animals expressing n5908[lin-29::gfp] (green channel), which fluorescently tags lin-29 at its endogenous locus, and nEx3004[MAB-10::mCherry] (red channel). Inset: Hypodermal seam cell nuclei. Scale bar, 40 μm. b DIC and confocal fluorescent images of hypodermal cells in 1-day-old wild-type, mab-10(tm2497), ddx-23(n5705) and mab-10(tm2497); ddx-23(n5705) adults expressing the reporter n5908[lin-29::gfp]. Images are representative of 5-8 animals per genotype. Scale bar, 40 μm. c RNA-Seq analysis of lin-29, mab-10, and ddx-23 mutant adults. The number of genes upregulated (black) and downregulated (gray) relative to wild-type adults are shown for the indicated genotypes. d Protein domain enrichment analysis of genes that are upregulated in lin-29, mab-10, and ddx-23 mutant adults, as identified by RNA-Seq analysis. Statistical tests were performed according to DAVID functional annotation tools^,, where Fisher’s exact test was used to determine P values and false discovery rates were calculated by the Benjamini–Hochberg procedure to correct for multiple testing. e A proposed model for how DDX-23-driven formation of MAB-10 nuclear foci controls LIN-29/MAB-10 co-regulated target gene repression in the adult hypoderm. DDX-23 binds to and enhances the partitioning of MAB-10 proteins into repressive foci, causing silencing of LIN-29/MAB-10 co-repressed target genes—including larval-specific and Hedgehog-related genes—and thereby defining adult cell identity (left panel). Loss-of-function mutations in ddx-23 reduce MAB-10 partitioning into LIN-29 repressive foci and upregulate LIN-29/MAB-10 co-repressed genes including larval-specific and Hedgehog-related genes, resulting in a failure to maintain adult cell identity (right panel). Color key for proteins are as follows: DDX-23 (purple), MAB-10 (blue), LIN-29 (red), chromatin factor (yellow). f Models for how the heterochronic pathway regulates cell differentiation and the juvenile-to-adult transition in C. elegans (upper panel) and mammals (lower panel). For example, LIN28 signaling in mammals (LIN-28 signaling in C. elegans) coordinates the expression of EGR (C. elegans LIN-29) and NAB (C. elegans MAB-10) proteins, both let-7 microRNA-dependently via TRIM71 (mammalian homolog of the C. elegans LIN-41) and let-7-independently via IKAROS (mammalian homolog of the C. elegans HBL-1) and GFI1. The formation of nuclear foci that contain NAB (MAB-10) proteins is facilitated by a DEAD-box helicase protein (DDX-23) and function to transcriptionally repress EGR (LIN-29) target genes including proliferation-related genes and hedgehog-related genes. The repression of proliferation-related genes leads to cellular differentiation, while Hedgehog signaling leads to luteinizing hormone LHβ expression and the onset of puberty. DDX-23 was previously described to play a role in the primary microRNA processing of let-7 microRNA (dotted lines), possibly providing a feedback regulatory loop in this pathway. Source data for (c) is provided as a Source Data file.