Regulated nuclear accumulation of a histone methyltransferase times the onset of heterochromatin formation in C. elegans embryos

Abstract

Heterochromatin formation during early embryogenesis is timed precisely, but how this process is regulated remains elusive. We report the discovery of a histone methyltransferase complex whose nuclear accumulation and activation establish the onset of heterochromatin formation in Caenorhabditis elegans embryos. We find that the inception of heterochromatin generation coincides with the accumulation of the histone H3 lysine 9 (H3K9) methyltransferase MET-2 (SETDB) into nuclear hubs. The absence of MET-2 results in delayed and disturbed heterochromatin formation, whereas accelerated nuclear localization of the methyltransferase leads to precocious H3K9 methylation. We identify two factors that bind to and function with MET-2: LIN-65, which resembles activating transcription factor 7-interacting protein (ATF7IP) and localizes MET-2 into nuclear hubs, and ARLE-14, which is orthologous to adenosine 5'-diphosphate-ribosylation factor-like 14 effector protein (ARL14EP) and promotes stable association of MET-2 with chromatin. These data reveal that nuclear accumulation of MET-2 in conjunction with LIN-65 and ARLE-14 regulates timing of heterochromatin domains during embryogenesis.

Fig. 1. Heterochromatin and H3K9me domains are established during embryogenesis.

(A) Timeline of C. elegans embryogenesis. Stages are color-coded: light green (<20 cells, pregastrula), green (21 to 200 cells, gastrula), and dark green (200 to 500 cells, late stage). Morphogenesis starts after the 500-cell stage and was not analyzed in this study. (B) Transmission electron micrographs of representative nuclei from WT embryos. Scale bars, 1 μm. (C) Survey of histone modifications (HMs). Representative single nuclei at designated embryonic stages stained for histones and DNA. Scale bars, 2 μm. DAPI, 4′,6-diamidino-2-phenylindole. (D) Quantitation of histone modifications normalized to total histone H3. Error bars denote SEM. a.u., arbitrary units.

Fig. 2. H3K9 methyltransferase MET-2/SETDB1 is required for heterochromatin formation.

(A) TEM of representative nuclei from WT or met-2 embryos. Scale bars, 1 μm. (B) TEM line-scan analysis for WT (black) or met-2 (magenta) nuclei and the SD. Line scan for a single nucleus is shown as an example. (C) Proportion of pixels within EDRs at the 200-cell stage. Error bars denote SEM. (D) Contrast of the cytosol in WT versus met-2 mutants and intensity thresholding (red) to define electron-dense structures. WT nuclei are in green and met-2 nuclei are in magenta.

Fig. 3. Nuclear accumulation of MET-2 determines the onset of H3K9 dimethylation.

(A) Embryos stained for MET-2::GFP (top) and HIS-72::mCherry (bottom) at designated stages of embryogenesis. Scale bars, 2 μm. (B and C) Quantitation of nuclear (B) and total (C) MET-2 at designated stages, normalized to HIS-72::mCherry. Error bars denote SEM. (D and E) Localization of 3xFLAG::MET-2 with a c-Myc NLS compared to 3xFLAG::MET-2 control (D) and corresponding H3K9me2 levels (E). We used histone H3 as a staining control. Note that H3 appears to be released during mitosis, but we focused on interphase cells. (F) Interphase nuclei showing H3K9me2 levels for the c-myc NLS construct compared to on-slide control embryos. (G) 3xFLAG::MET-2 line scans in pregastrula embryos (one- to four-cell stage) with (red) or without (gray) the c-myc NLS. Average of line scans across multiple nuclei is shown, and error bars denote SEM. (H) H3K9me2 levels for the NLS construct (red), normalized to H3 and to WT embryos at the 51- to 100-cell stage (gray). Error bars denote SEM.

Fig. 4. MET-2 and two conserved binding partners form concentrated nuclear hubs at gastrulation.

(A) Whole embryos stained with antibodies against methylated H3K9 and pan-histones in WT versus met-2, lin-65, or arle-14 mutants. We used H3K9me2 (Kimura 6D11) antibodies. Scale bars, 2 μm. (B) H3K9me levels in mutant embryos normalized to on-slide control embryos and to pan-histone. Error bars denote SEM. (C) Representative single nuclei showing H3K9me3 (red) and DAPI (blue) in WT versus met-2, lin-65, or arle-14 mutants. Scale bars, 2 μm. (D) H3K9me3 line-scan analysis for WT (green), met-2 (magenta), lin-65 (orange), arle-14 (yellow) nuclei, and the SD. Line scan for a single nucleus is shown as an example. (E) Expression of CEREP4 and CEMUDR1 repeat RNAs by real-time quantitative polymerase chain reaction (qPCR) in WT versus met-2, lin-65, or arle-14 mutant embryos. (F) Definition of hubs by intensity thresholding. (G) Single nuclei showing 3xFLAG::MET-2, LIN-65::3xFLAG, and ARLE-14 colocalization in concentrated protein hubs and exclusion of activating mark H3K4me3. Scale bars, 2 μm. (H to J) Representative single nuclei showing 3xFLAG::MET-2 (H), LIN-65::3xFLAG (I), and ARLE-14 (J) localization at different embryonic stages. Scale bars, 2 μm. Quantitation of signal intensity in nuclear hubs (brown), nuclear regions excluding hubs (“nonhub,” dark gray), and cytosol (light gray). Error bars denote SEM. (K) PLA signal showing interactions between MET-2/LIN-65 and MET-2/ARLE-14. (L and M) Percentage of nuclear PLA dots per total dots in embryos for MET-2/LIN-65 (L) and MET-2/ARLE-14 (M) interactions. Error bars denote SEM.

Fig. 5. ARLE-14 promotes chromatin association of MET-2.

(A) Whole embryos showing the distribution of MET-2::GFP in WT versus arle-14 mutants. Scale bars, 2 μm. (B) Line-scan analysis showing the mean MET-2::GFP intensity across embryonic cells in WT versus arle-14 mutant embryos. Average of line scans across multiple nuclei is shown, and error bars denote SEM. (C) ARLE-14 antibody staining in WT versus met-2, lin-65, and arle-14 mutants. (D) ARLE-14 line-scan quantitation. Average of line scans across multiple nuclei is shown, and error bars denote SEM. (E and F) H3K9me2 ChIP-seq (logLR) track in WT versus arle-14 mutant embryos (chromosome III and loci: rep-1, Y22D7AL.7, and grl-16). Five WT ChIP experiments and 10 arle-14 mutant ChIP experiments were pooled in for sequencing, and results are not quantitative. (G) H3K9me2 ChIP-qPCR in WT versus arle-14 mutants before combining parallel ChIP experiments together for sequencing. Error bars denote SEM. (H) MET-2::GFP ChIP-qPCR for WT (green) versus arle-14 (yellow) mutant embryos. Inset shows H3K4me3 ChIP-qPCR as a control. The y axis shows normalized %input values for both WT and mutants, where the %input value at the H3K4me3-positive region was set to 1 for both genotypes. %input for MET-2::GFP ChIP was adjusted accordingly. Error bars denote SEM for n = 3 experiments. (I and J) Acquisition of H3K9me2 in WT versus arle-14 mutants during embryogenesis, with an H3 costain (I) and quantitation normalized to H3 (J). Error bars denote SEM. (K) Amino acid sequence alignment of worm ARLE-14 to the human ARL14EP domain using ClustalX2. Color scheme: blue, hydrophobic; red, positive charge; magenta, negative charge; green, polar; pink, cysteines; orange, glycines; yellow, prolines; cyan, aromatic. (L) Protein interaction map for human ARL14EP from STRING database (https://string-db.org/). The red box highlights interactions with human SETDB1/2.

Fig. 6. LIN-65 is rate-limiting for H3K9me2 and nuclear MET-2 during embryogenesis.

(A) Distribution of MET-2::GFP in WT versus lin-65 mutants versus no-GFP WT strain. Scale bars, 2 μm. Note that this H3 antibody detects mostly cytosolic histone H3 during mitosis. (B) Line-scan analysis across embryonic nuclei shows mean MET-2::GFP intensity in WT (green) versus lin-65 (pink) mutants versus no-GFP control (“−,” gray). Average of line scans across multiple nuclei is shown, and error bars denote SEM. (C and E) H3K9me2 and H3K9me3 levels in embryos with a single-copy ZEN-4::GFP (“WT,” on-slide control) or progeny of lin-65(+/−) heterozygous mothers identified by HIS-72::mCherry at designated stages of embryogenesis. (D and F) Quantitation of H3K9me2 and H3K9me3 levels from WT (gray) or lin-65(+/−) (purple) offspring. Error bars denote SEM. (G) Domain architecture from HMMER (www.ebi.ac.uk/Tools/hmmer/) for human ATF7IP, fly Windei (WDE), and worm LIN-65 showing disordered (purple), coiled-coil (pink), and β-sandwich fold (violet) regions. (H) In early embryos, MET-2 (gray), LIN-65 (red), and ARLE-14 (yellow) are enriched in the cytosol, and there is little H3K9me2 or heterochromatin (light green). As embryos mature, MET-2 and interactors gradually accumulate in nuclei, form concentrated nuclear hubs, and deposit H3K9me2. MET-2–dependent H3K9me is required to generate heterochromatin domains (compacted, dark green).