Germ Granules Prevent Accumulation of Somatic Transcripts in the Adult Caenorhabditis elegans Germline

Abstract

The germ cells of multicellular organisms protect their developmental potential through specialized mechanisms. A shared feature of germ cells from worms to humans is the presence of nonmembrane-bound, ribonucleoprotein organelles called germ granules. Depletion of germ granules in Caenorhabditis elegans (i.e., P granules) leads to sterility and, in some germlines, expression of the neuronal transgene unc-119::gfp and the muscle myosin MYO-3 Thus, P granules are hypothesized to maintain germ cell totipotency by preventing somatic development, although the mechanism by which P granules carry out this function is unknown. In this study, we performed transcriptome and single molecule RNA-FISH analyses of dissected P granule-depleted gonads at different developmental stages. Our results demonstrate that P granules are necessary for adult germ cells to downregulate spermatogenesis RNAs and to prevent the accumulation of numerous soma-specific RNAs. P granule-depleted gonads that express the unc-119::gfp transgene also express many other genes involved in neuronal development and concomitantly lose expression of germ cell fate markers. Finally, we show that removal of either of two critical P-granule components, PGL-1 or GLH-1, is sufficient to cause germ cells to express UNC-119::GFP and MYO-3 and to display RNA accumulation defects similar to those observed after depletion of P granules. Our data identify P granules as critical modulators of the germline transcriptome and guardians of germ cell fate.

Keywords: Caenorhabditis elegans; GLH-1; P granules; PGL-1; germ-soma antagonism; germline.

Figure 1

Method to profile transcripts from PG(−) gonads. PG(−) is shorthand for P granule-depleted. (A) Experimental scheme and timeline to analyze the transcriptomes of control and PG(−) gonads from L4 larval, day 1 adult, and day 2 adult stages of worms grown at 24°. (B) Seven different samples were collected (three control, and four P-granule RNAi), all in triplicate. Throughout this report, the samples are color coded: control samples are in light blue, L4 and day 1 adult PG(−) UNC-119::GFP(−) samples are in pink, day 2 adult PG(−) UNC-119::GFP(−) samples are in red, and day 2 adult PG(−) UNC-119::GFP(+) samples are in green.

Figure 2

Large, progressive changes in transcript accumulation are observed in adult PG(−) gonads but not in L4 larval gonads. (A) Volcano plots comparing PG(−) gonads to control gonads at different time points. Dashed lines indicate the significance cut off of q = 0.05 (horizontal line) and a twofold change in transcript level (vertical lines). The numbers of significantly misregulated genes in each quadrant are shown at the top of each plot. (B) Venn diagrams showing the overlap of genes down- or upregulated in day 1 adult PG(−) UNC-119::GFP(−) gonads (pink), day 2 adult PG(−) UNC-119::GFP(−) gonads (red), and day 2 adult PG(−) UNC-119::GFP(+) gonads (green). The overlap of all three samples is shaded in orange. (C) PCA of all seven samples [PG(−) and control samples] across two principal components (PC1 and PC2).

Figure 3

PG(−) gonads downregulate germline genes and upregulate sperm and somatic genes. (A) Box plots of normalized RNA-seq reads from ubiquitous genes, genes with germline-enriched expression, genes with germline-specific expression, sperm genes, and soma-specific genes in control and PG(−) gonads at different stages. Each box extends from the 25th to the 75th percentile of read values. The whiskers extending from each box indicate the 2.5th and the 97.5th percentiles. Wedges around the median indicate 95% confidence intervals for the medians. (B–E) UCSC genome browser tracks showing the abundance of RNA-seq reads from control and PG(−) gonad samples across (B) the ubiquitous gene ama-1; (C) germline-specific genes gld-1 and csr-1; (D) sperm-specific genes msp-81, -10, and -56, nspd-2, and ssp-10; and (E) soma-specific genes let-2, cebp-1, and aqp-2. The sample names in the box plots and the genome browser tracks are abbreviated with L4 denoting L4 larvae, Ad1 denoting day 1 adults, and Ad2 denoting day 2 adults. Additionally, UNC(−) denotes no expression and UNC(+) denotes expression of the unc-119::gfp transgene. The box plots and browser tracks are color coded as described for Figure 1.

Figure 4

Soma-specific transcripts accumulate in PG(−) germlines. (A) smFISH analysis of cebp-1 transcripts (red) in dissected day 1 and day 2 adult control gonads, a day 1 adult PG(−) gonad, and a day 2 adult PG(−) UNC-119::GFP(+) gonad. (B) smFISH analysis of aqp-2 transcripts (red) in a day 2 adult control gonad and a day 2 adult PG(−) UNC-119::GFP(+) gonad. For each gonad in (A) and (B), UNC-119::GFP signal is in green and DAPI-stained DNA is in blue. The distal half of each gonad is shown as a z-stack projection, with the distal tip indicated by *. Boxed regions are expanded in the right-most panels. The zoomed-in regions are projections of three slices taken from the middle section of the gonad.

Figure 5

Day 2 adult PG(−) UNC-119::GFP(+) gonads upregulate genes involved in neuronal processes and downregulate genes involved in germline processes. (A) Volcano plot comparing day 2 adult PG(−) UNC-119::GFP(+) gonads to day 2 adult PG(−) UNC-119::GFP(−) gonads. Dashed lines indicate the significance cut off of q = 0.05 (horizontal line) and a twofold change in transcript level (vertical lines). The numbers of significantly misregulated genes in each quadrant are shown at the top of the plot. (B) Gene ontology analysis of the 2268 upregulated genes and 692 downregulated genes in day 2 adult PG(−) UNC-119::GFP(+) gonads compared to day 2 adult PG(−) UNC-119::GFP(−) gonads. The gene ontology processes are organized based on their P-values, and the numbers of genes in each ontology category are shown in parentheses. (C) Gonads from day 2 adult control and day 2 adult PG(−) UNC-119:GFP(+) worms were dissected and immunostained for UNC-64 (red), GFP (green), and DNA (blue). The distal half of each gonad is shown as a z-stack projection, with the distal tip indicated by *. Boxed regions are expanded in the right-most panels. The zoomed-in regions are projections of three slices taken from the middle section of the gonad.

Figure 6

Acquisition of neuronal fate in PG(−) germ cells is accompanied by loss of germ cell markers. (A) Gonads from day 2 adult control and day 2 adult PG(−) UNC-119::GFP(+) worms were dissected and immunostained for REC-8 (red), GFP (green), and DNA (blue). (B) Quantification of UNC-119::GFP signal relative to REC-8 signal in germ cells from control and PG(−) gonads. (C) Same as in (A) except gonads were stained for HTP-3 (red). (D) Same as in (B) except gonads were quantified for UNC-119::GFP signal relative to HTP-3 signal. (E) smFISH analysis of lag-1 transcripts (red) in a dissected day 2 adult control gonad and a day 2 adult PG(−) UNC-119::GFP(+) gonad. For (A), (C), and (E), the distal half of each gonad is shown as a z-stack projection, with the distal tip indicated by *. Boxed regions are expanded in the right-most panels. The zoomed-in regions are projections of three slices taken from the middle section of the gonad, except for the (C) control HTP-3 gonad, which shows the top section of the gonad. In the expanded panels in (C), the red channel of the PG(−) gonad was increased to better show relative HTP-3 levels.

Figure 7

Germlines lacking PGL-1 or GLH-1 also misexpress somatic markers. (A) Percent of germlines expressing the neuronal marker UNC-119::GFP (green bars) in glh-1 single mutants, glh-1 glh-4 double mutants, pgl-1 single RNAi, and pgl-1; pgl-3 double RNAi animals containing the unc-119::gfp transgene. Of the germlines that expressed UNC-119::GFP, a fraction also expressed the muscle marker MYO-3 (green and red striped bars). (B) Volcano plots comparing glh-1 (gold) or pgl-1 (blue) single mutant gonads to wild-type gonads. All animals lacked the unc-119::gfp transgene. Dashed lines indicate the significance cut off of q = 0.05 (horizontal line) and a twofold change in transcript levels (vertical lines). The numbers of significantly misregulated genes in each quadrant are shown at the top of the plot. (C) Venn diagrams showing the overlap of genes down- or upregulated in glh-1 and pgl-1 single mutant gonads.

Figure 8

Models of P-granule function. P granules reside over nuclear pores, where they intercept stochastically misexpressed soma-specific mRNAs (red mRNAs), while allowing germline-appropriate mRNAs (green mRNAs) to transit to the cytoplasm for translation. (Left panel) In model A, P granules prevent the translation of somatic mRNAs. (Middle panel) In model B, P granules slice somatic mRNAs into small RNAs, which in turn induce heritable silencing of somatic genes. (Right panel) In the absence of P granules, somatic mRNAs accumulate and are translated into proteins (red) that specify somatic identity and antagonize germline identity.