Drosophila germ granules are structured and contain homotypic mRNA clusters

Abstract

Germ granules, specialized ribonucleoprotein particles, are a hallmark of all germ cells. In Drosophila, an estimated 200 mRNAs are enriched in the germ plasm, and some of these have important, often conserved roles in germ cell formation, specification, survival and migration. How mRNAs are spatially distributed within a germ granule and whether their position defines functional properties is unclear. Here we show, using single-molecule FISH and structured illumination microscopy, a super-resolution approach, that mRNAs are spatially organized within the granule whereas core germ plasm proteins are distributed evenly throughout the granule. Multiple copies of single mRNAs organize into 'homotypic clusters' that occupy defined positions within the center or periphery of the granule. This organization, which is maintained during embryogenesis and independent of the translational or degradation activity of mRNAs, reveals new regulatory mechanisms for germ plasm mRNAs that may be applicable to other mRNA granules.

Figure 1. Germ plasm proteins occupy the same space within a granule.

(a) A confocal image of an embryo expressing VasaGFP. (b,c) Images of embryos expressing VasaKuOr (red) and OskGFP (green) acquired with a widefield epifluorescence microscope, and a confocal image of an embryo expressing VasaGFP transgene (green) and immunostained against Aub (red) or Tud (red). In all panels embryos fixed at 0–1 h AEL were used. (d) PCC showing co-localization between Vasa, Osk, Aub and Tud. Nine, four and eight embryos were analysed for OskGFP/VasaKuOr pair, VasaGFP/Aub pair and VasaGFP/Tud pair, respectively. An average±s.e.m. is shown. Scale bar in (a) (embryo) 50 μm, in (a) (blow-up) and (b) 10 μm, in (c) 1 μm.

Figure 2. Germ plasm-enriched mRNAs occupy distinct positions within the VasaGFP granule.

(a,b) SIM images of early Drosophila embryos (0–1 h AEL) expressing VasaGFP (green) and stained with CALFluor590-labelled smFISH probes, targeting respectively cycB, nos, pgc, gcl or osk (red). (c) Per cent of overlapping VasaGFP granules with mRNAs and VasaKuOr granules with OskGFP granules. An average±s.e.m. is shown. (d) Ratio between the number of cycB, nos, pgc, gcl and osk mRNA particles and VasaGFP granules obtained in c. For each embryo a ratio between the number of mRNA particles and VasaGFP was calculated. An average±s.d. is shown. (e) Per cent of VasaGFP or VasaKuOr area overlapping with cycB, nos, pgc or gcl mRNAs or OskGFP, respectively. Overlap measures the area of VasaGFP granule occupied by a corresponding overlapping mRNA particle. mRNA particles and granules in c were analysed. An average±s.e.m. is shown. (f) Size of localized mRNA particles and VasaGFP granules measured in pixels where pixel size X=Y=56 nm. An average±s.e.m. is plotted. nos, gcl and osk particles were smaller than cycB particles (**t-test, two-tailed P=0.01, ***t-test, two-tailed P=0.001). osk particles were smaller than pgc particles (*t-test, two-tailed P=0.03). (g) A SIM image of localized pgc mRNA labelled with a mix of Alexa488 (green) and CALFluor590 (red) smFISH probes. (h) A widefield epifluorescence image of ccr4 mRNA (red) in embryo expressing VasaGFP (green). ccr4 is not enriched at the posterior pole. (i) Co-localization of mRNAs within the VasaGFP granules, OskGFP granules within the VasaKuOr granules and Alexa488-labelled pgc mRNA within CalFluor590-labelled pgc mRNA determined by measuring the distance between the center of the VasaGFP granule and the center of the overlapping mRNA and by determining the PCC(Costes) (Methods section). An average±s.e.m. is plotted. Scale bar in a, h (right panel) 10 μm, in b, g and h (left panel) 1 μm.

Figure 3. gcl but not osk is a constituent of the germ granule.

(a) Two colour smFISH and SIM were used for detection of co-localized transcripts. (b) Per cent of overlapping mRNA particles as described in Fig. 2c. An average±s.e.m. is shown. (c) Per cent of overlapping osk mRNA particles with cycB, nos, gcl and pgc mRNA particles. An average±s.e.m. is shown. (d) Pairwise mRNA co-localization. Distance measurements and PCC(Costes) were performed as described in Fig. 2i. An average±s.e.m. is plotted. (e) Co-localization of cycB, nos, pgc, gcl and osk mRNA with OskGFP granules as quantified by PCC(Costes). An average±s.e.m. is plotted. (f,g) Venn diagrams showing concurrent overlapping of VasaGFP with cycB and nos and VasaGFP with cycB and gcl. Six and seven embryos expressing VasaGFP and double labeled to detect cycB and nos or cycB and gcl were analysed, respectively (Supplementary Fig. 3j). The numbers within Venn circles indicate number of VasaGFP granules and cycB, nos and gcl mRNA particles. An average±s.e.m. is shown. Scale bar in a 1 μm.

Figure 4. Localized mRNAs form homotypic rather than mixed clusters.

(a) SIM and smFISH were used to detect nos (green) and pgc (red) mRNAs located ventrally and at the posterior of the embryo. (b) The concentration (nM) of nos, pgc and gcl mRNA found in the embryo (black bars) and localized at the posterior pole (red bars) determined by smFISH (Methods section). The localization efficiency of mRNAs at the posterior pole is indicated above the bars (Supplementary Fig. 5c,f–h, Supplementary Table 3 Methods section). Eleven embryos/mRNA were analysed. An average±s.e.m. is shown. (c) Co-localization of nos mRNAs and gcl mRNAs with pgc mRNAs ventrally and at the posterior of an embryo. An average±s.e.m. is shown. (d,e) Multiple nos or pgc mRNAs occupy a single nos or pgc mRNA cluster at the posterior. Ventrally nos and pgc were mostly found as single mRNAs. (f) A SIM image of nos (green) and pgc (red) mRNA with an accompanying spatial map of co-localization of homotypic nos (green) and pgc (red) mRNA clusters. The sub-pixel position of nos and pgc clusters and corresponding number of nos and pgc mRNAs per cluster was determined with Airlocalize (Methods section). (g) A confocal image of endogenous (wild type (WT)) nos mRNA (green) and the chimeric mRNA with GFP ORF and nos 5′ and 3′ UTRs (red) co-localizing at the posterior pole. Below the image is a depiction of both mRNAs labelled with either red or green smFISH probes. (h) A confocal image of WT pgc mRNA (red) and the chimeric mRNA with GFP ORF and nos 5′ and 3′ UTRs (green) co-localizing at the posterior pole. (i) Co-localization of WT nos mRNA with the chimeric GFP-nos 5′, 3′ UTR mRNA (four embryos), of WT pgc labelled with Alexa488 probes and WT pgc labelled with CALFluor590 probes (two embryos), of WT pgc and the chimeric GFP-nos 5′, 3′ UTR mRNA (six embryos) and WT pgc mRNA and WT nos mRNA (six embryos) quantified by PCC(Costes). (j) Using triangulation a 3D model of an average VasaGFP (pink) granule with localized cycB (blue), nos (green), pgc (yellow) and gcl (red) is shown (Methods section). Scale bar in a 5 μm, in f–h 1 μm.

Figure 5. The position of mRNAs within the granule is independent of the translational or degradation onsets.

(a) Embryos expressing VasaGFP transgene were fixed after 0–1 h AEL (black bars) or 1–1.5 h AEL (red bars) and smFISH for cycB, nos, pgc and gcl was performed. Black bars are PCC(Costes) measurements from Fig. 2i. For the PCC(Costes) analysis performed on 1–1.5 h AEL embryos (red bars), eight, nine, six and five embryos were analysed for the cycB mRNA/VasaGFP, nos mRNA/VasaGFP, pgc mRNA/VasaGFP and gcl mRNA/VasaGFP pairs, respectively. An average±s.e.m. is plotted. (b) SIM images of 0–1 h AEL w¹¹¹⁸ embryo localizing pgc (green) and nos (red) mRNAs at the posterior pole. (c) SIM images of 1–1.5 h AEL w¹¹¹⁸ embryo of pgc (green) and nos (red) mRNAs surrounding PGC nuclei. (d) w¹¹¹⁸ embryos were fixed 0–1 h AEL (black bars) or 1–1.5 h AEL (red bars). A two colour smFISH was performed to detect pgc mRNA and either cycB, gcl or osk mRNA. Alexa488-labelled RNA probes were used to detect pgc mRNA and CALFluor590-labelled probes were used to detect cycB, gcl or osk mRNA (b,c). PCC(Costes) analysis was performed as described above. For the PCC(Costes) analysis performed on 0–1 h AEL embryos, 7, 7, 13 and 7 embryos were analysed for the cycB mRNA/pgc mRNA, nos mRNA/pgc mRNA, gcl mRNA/pgc mRNA and osk mRNA/pgc mRNA pairs, respectively. For the PCC(Costes) analysis performed on 1–1.5 h embryos, 7, 8, 10 and 7 embryos were analysed for the cycB mRNA/pgc mRNA, nos mRNA/pgc mRNA, gcl mRNA/pgc mRNA and osk mRNA/pgc mRNA pairs, respectively. An average±s.e.m. is plotted. **t-test, two-tailed P<0.0001, *t-test, two-tailed P=0.01. (e) In 0–1 h AEL embryos unlocalized cycB and gcl mRNAs are stable until the activation of zygotic genome∼2.5 h AEL (Zygotic) while unlocalized nos and pgc are unstable and decay before and during the activation of zygotic genome (Maternal & Zygotic). PCC(Costes) values were obtained in Fig. 2i. Scale bar in b,c 10 μm.