A global chromatin compaction pathway that represses germline gene expression during starvation

Abstract

While much is known about how transcription is controlled at individual genes, comparatively little is known about how cells regulate gene expression on a genome-wide level. Here, we identify a molecular pathway in the C. elegans germline that controls transcription globally in response to nutritional stress. We report that when embryos hatch into L1 larvae, they sense the nutritional status of their environment, and if food is unavailable, they repress gene expression via a global chromatin compaction (GCC) pathway. GCC is triggered by the energy-sensing kinase AMPK and is mediated by a novel mechanism that involves the topoisomerase II/condensin II axis acting upstream of heterochromatin assembly. When the GCC pathway is inactivated, then transcription persists during starvation. These results define a new mode of whole-genome control of transcription.

Figure 1.

Developmental timing of chromatin compaction in C. elegans primordial germ cells.(A) Germ cell nuclei in living WMM1 embryos of various stages were imaged for chromatin compaction using an mCherry-tagged histone H2B. Scale bar represents a length of 2 µm (n = 8). (B) Schematic showing the two regions of the nucleus used for quantification in subsequent figures. (C) Quantification of the data presented in A. Vertical axis represents the percentage of pixels in the inner compartment described in B, and the horizontal axis presents germ cell nuclei at different stages of development. Error bars represent one standard deviation. P values were obtained using the Student’s t test. (D) Germ cell nuclei of N2 embryos and larvae at various developmental stages were fixed and stained for P-granules (white), RNAPII (pSer2; red), and DNA (blue). Representative images are shown, with white arrows pointing to P3 in the first panel and Z2/Z3 in the remaining panels. The number of samples examined ranged from 10 (early embryos) to 22 (starved L1s). See Table S1 for a summary of pSer2 signal intensities across the entire dataset.

Figure S1.

pSer2 antibody reactivity requires the presence of phosphorylated RNA polymerase II. Embryos were fixed and stained for P-granules (green), pSer2 (red), and DNA (blue). Both ama-1 RNAi treatment (which depletes RNAPII) and calf intestine phosphatase treatment result in the loss of pSer2 signal. Thus, the pSer2 signal is dependent on the presence of phosphorylated RNAPII. Scale bar represents a length of 2 µm.

Figure 2.

The TOP-2/condensin II axis and heterochromatin pathway are required for Z2/Z3 chromatin compaction.(A) Strain WMM1 was treated with either control or capg-2 RNAi, and F1s born from these animals were starved upon hatching. Z2/Z3 nuclei of the starved L1s were then imaged for chromatin compaction. (n = 20). (B) Same as A except samples were treated with cec-4 RNAi and not capg-2 RNAi. (C) Chromatin compaction was quantified in starved L1s born from strain WMM1 that was treated with either capg-2, kle-2, top-2, hpl-2, cec-4, or mys-1 RNAi, as indicated. Each gene-targeted RNAi was accompanied by a control RNAi; thus, the data are presented as pairs of control and gene-targeted samples. Error bars represent one standard deviation (n = 20). (D) Strain WMM1 was treated with either control or capg-2 RNAi, and F1s born from these animals were examined. Z2/Z3 nuclei of the new L1s were imaged for chromatin compaction. (n = 20). (E) Same as D except samples were treated with cec-4 RNAi and not capg-2 RNAi. (F) Chromatin compaction was quantified in new L1s born from strain WMM1 that was treated with either capg-2, top-2, hpl-2, and cec-4 RNAi. Data are presented as in C. Error bars represent one standard deviation (n = 20). Scale bar represents a length of 2 µm.

Figure S2.

The TOP-2/condensin II axis and heterochromatin pathway are required for Z2/Z3 chromatin compaction.(A) Samples were treated as in Fig. 2 A except the indicated RNAi was used. (B) Samples were treated as in Fig. 2 D except the indicated RNAi was used. (C) Starved N2 and RB2301 (cec-4^−/−) L1s were stained with Hoechst 33342 DNA dye and live-imaged for Z2/Z3 chromatin compaction. Quantification is shown below. Error bars represent one standard deviation (n = 20). (D) Somatic cells in the vicinity of Z2/Z3 in starved L1s, born from strain WMM1 that was treated with either control or cec-4 RNAi, were imaged for chromatin compaction. Quantification is shown below. Error bars represent one standard deviation (n = 20). Scale bar represents a length of 2 µm.

Figure S3.

Widespread H3K9 methylation is observed as chromatin compaction starts in C. elegans PGCs.(A)C. elegans embryos at different embryonic stages were fixed and stained for P-granules (green), H3K9me3 (red), and DNA (blue). Representative whole-embryo images are shown. (B) Same as A except only Z2/Z3 (germ) or a neighboring somatic cell (soma) is shown. The developmental stage of the embryo from which the image was taken is indicated (n = 20). See Table S1 for a summary of signal intensities. Scale bar represents a length of 2 µm.

Figure 3.

A TOP-2/condensin II–dependent increase in heterochromatin marks coincides with chromatin compaction in Z2/Z3.(A) Wild-type starved L1s and starved L1s that were defective for methyltransferases (mutants for met-2, set-25, set-32, and F1s from animals treated with met-2/set-25 double RNAi) were fixed and stained for P-granules (white), H3K9me2 (green), H3K9me3 (red), and DNA (blue). (n = 40). See Table S1 for a summary of signal intensities. (B) L1s were either starved or fed for a varying amount of time. Samples were then fixed and stained for P-granules (white), H3K9me3 (red), H3K9me2 (green), and DNA (blue). Representative images are shown. (n = 40). See Table S1 for a summary of signal intensities. (C) Starved L1s, born from strain N2 treated with either control RNAi or capg-2 RNAi, were fixed and stained for P-granules (white), H3K9me2 (green), H3K9me3 (red), and DNA (blue). Representative images are shown (n = 40). See Table S1 for a summary of signal intensities. (D) Starved L1s, born from strain N2 treated with either control or top-2 RNAi, were fixed and stained for H3K9me3 (red) and DNA (blue). Representative images are shown (n = 40). See Table S1 for a summary of signal intensities. (E) Worms that express HPL-2::mKate were optionally treated with control and capg-2 RNAi. Live embryos were extracted and were imaged for HPL-2 signal. The white star identifies Z2/Z3. Representative images are shown (n = 20). (F) Quantification of HPL-2::mKate signal from the images taken in E. Error bars represent one standard deviation. Scale bar represents a length of 2 µm.

Figure S4.

Both H3K9 methyltransferases, MET-2 and SET-25, are needed for chromatin compaction in starved L1s. New and starved L1s born from strain WMM1 treated with either control RNAi, met-2 RNAi, or set-25 RNAi or met-2/set-25 double RNAi were used. Z2/Z3 from the L1s were imaged for chromatin compaction. Representative images are shown. Compaction was also quantified and shown below images. Error bars represent one standard deviation (n = 20). Scale bar represents a length of 2 µm.

Figure 4.

Aberrant germline transcription is observed when GCC components are depleted.(A) Starved L1s, born from strain N2 that had been treated with either control RNAi or capg-2 RNAi, were fixed and stained for P-granules (P-gran.; white), DNA (blue), and RNAPII (pSer2; green). The number of pSer2-positive samples is shown below (n = 20). (B) Starved L1s, born from strain N2 that had been treated with either control RNAi, cec-4 RNAi, or met-2/set25 RNAi, were fixed and stained for P-granules (green), DNA (blue), and activated RNAPII (pSer2; red). The number of samples that were positive for pSer2 signal is shown below (n = 40). (C) Starved L1s, born from strain N2 that had been treated with either control RNAi, cec-4 RNAi, or met-2/set25 RNAi, were used in this experiment. In situ HCR was performed by probing for wago-1 (red) and cgh-1 (white). DNA was visualized using Hoechst 33342 dye (blue), and xnd-1 (green) was used to identify Z2/Z3. The white dashed lines correspond to the nuclei. Percentage of samples positive for each gene is shown on the right (n = 40). (D) Starved L1s, born from strain N2 that had been treated with either control or top-2 RNAi were used. HCR for xnd-1 (green), ifet-1 (red), and car-1 (white) was performed. DNA was visualized using Hoechst 33342 dye (blue). Percentage of samples positive for each gene is shown on the right (n = 25). Scale bar represents a length of 2 µm.

Figure 5.

GCC pathway and its activator AMPK are required for the nascent germline to survive nutritional stress.(A) Chromatin compaction was quantified in starved L1s born from strain WMM1 that was treated with either control or aak-1/2 RNAi. Quantification is shown below. Error bars represent one standard deviation (n = 20). (B) Chromatin compaction was quantified in new L1s born from strain WMM1 that was treated with either control or aak-1/2 RNAi. Quantification is shown below. Error bars represent one standard deviation (n = 20). (C) Starved L1s, born from strain N2 that had been treated with either control RNAi or aak-1/2 RNAi, were fixed and stained for P-granules (P-gran.; white), DNA (blue), and RNAPII (pSer2; red). The number of pSer2-positive samples is shown below (n = 20). Scale bar represents a length of 2 µm.

Figure 6.

Summary of events that comprise a genome activation-repression-reactivation cycle in the developing C. elegans germline. Please see main text for details.