A role for SOX9 in post-transcriptional processes: insights from the amphibian oocyte

Abstract

Sox9 is a member of the gene family of SOX transcription factors, which is highly conserved among vertebrates. It is involved in different developmental processes including gonadogenesis. In all amniote species examined thus far, Sox9 is expressed in the Sertoli cells of the male gonad, suggesting an evolutionarily conserved role in testis development. However, in the anamniotes, fishes and amphibians, it is also expressed in the oocyte but the significance of such an expression remains to be elucidated. Here, we have investigated the nuclear localization of the SOX9 protein in the oocyte of three amphibian species, the urodelan Pleurodeles waltl, and two anurans, Xenopus laevis and Xenopus tropicalis. We demonstrate that SOX9 is associated with ribonucleoprotein (RNP) transcripts of lampbrush chromosomes in an RNA-dependent manner. This association can be visualized by Super-resolution Structured Illumination Microscopy (SIM). Our results suggest that SOX9, known to bind DNA, also carries an additional function in the posttranscriptional processes. We also discuss the significance of the acquisition or loss of Sox9 expression in the oocyte during evolution at the transition between anamniotes and amniotes.

Figure 1

Immunodetection of SOX9 in the oocyte of X. tropicalis, X. laevis and P. walt. Western blot analysis of SOX9 in nuclear and cytoplasmic extracts using the polyclonal anti-Cter-SOX9 antibody and the monoclonal anti-tubulin antibody. (N) corresponds to 25 GVs from X. tropicalis (X.t), 20 from X. laevis (X.l) and 15 from P. waltl (P.w). (N’) nuclear extract without nuclear envelope. (N”) 15 GVs from X. laevis. (Cy) cytoplasms from 10X. tropicalis oocytes and from 5X. laevis oocytes. The anti-Cter- SOX9 antibody recognized a major polypeptide around 68-kDa in the nuclear (N and N’) and cytoplasmic (Cy) extracts of the three species (black double arrowheads) and an additional one at 48 kDa (black arrowhead). The control anti-tubulin antibody recognized a 50-kDa polypeptide only in the cytoplasmic extracts (empty arrowheads).

Figure 2

Subnuclear localization of the SOX9 protein in the GV of X. tropicalis, X. laevis and P. waltl. Fluorescent and corresponding phase contrast micrographs of nuclear spreads showing one LBC of X. tropicalis and X. laevis from stage V-VI oocyte, and part of a P. waltl LBC from stages II and V-VI oocyte. Nuclear spreads were immunostained with the anti-Cter-SOX9 antibody (green, Alexa 488 IgG) and counterstained with Hoechst (red). The Hoechst dye stained the chromosome DNA axis and not that of the decondensed lateral loops. The antibody labeled the LBCs, and the nuclear bodies either attached to them (empty arrowheads) or free in the nucleoplasm (full arrowheads). Note that the immunostaining of lateral loops was very clear in P. waltl LBCs from stage II oocyte, because of their extension and density. The intensity of fluorescence in the different experiments was normalized with respect to their relevant controls with the secondary antibody (tagged with Alexa fluor 488) alone. Wide field Leica microscope. Scale bar for all micrographs: 10 μm.

Figure 3

Targeting of the newly-synthetized Xt-SOX9-GFP protein to LBCs and nuclear bodies. (A) Capped, in vitro-synthetized transcripts from the pcDNA Xt-sox9- CT-GFP (Sox9) vector, or the control pcDNA NLS-CT-GFP (NLS) vector, were injected into the cytoplasm of stage IV-V oocytes of X. laevis. The expressed proteins (Xt-SOX9-GFP or NLS-GFP) were detected 18 hours later, either directly on nuclear spreads by fluorescence microscopy or indirectly on immunoblots using antibodies against the SOX9 or GFP proteins. (B) Immunoblots of nuclear extracts from injected oocytes with the sox9-CT-GFP (Sox9) or NLS-CT-GFP (NLS) transcripts. Each lane corresponds to 15 GVs. kDa: molecular weight markers. A polypeptide (*) around 80 kDa detected with the anti-Nter or Cter- SOX9 antibodies and the anti-GFP antibody in the Sox9-CT-GFP injected oocytes corresponded to Sox9-GFP protein. In the NLS-CT-GFP injected oocytes, the anti-GFP antibody detected a 30 kDa polypeptide (°) corresponding the NLS-GFP protein. The polypeptide of ca. 68 kDa (arrows) corresponds to the endogenous SOX9 protein. The dashed lines separate the lanes which were cut from the original immunoblots shown in the supplementary figure S5. (C) (b, b’) Direct detection of GFP-SOX9 and GFP-NLS proteins on nuclear spreads from the injected oocytes. The GFP-SOX9 protein targeted the LBCs and the free nuclear bodies (arrowhead). A much lower level of GFP-NLS was detected on the LBCs. (c, c’) immunostaining of the nuclear spreads using the anti-Cter-SOX9 antibody. (d, d’) merged images (b, b’ and c, c’) indicating that the staining pattern of the Xt-SOX9-GFP was similar to that using the anti-Cter-SOX9 antibody. Wide field Leica microscope. Scale bar for all micrographs: 10 μm. (D) Quantification of GFP fluorescence density on LBCs with the Otsu’s method. GFP-fluorescence density over the GFP-SOX9 LBCs was significantly higher than that of the control GFP-NLS LBCs (Student’s test, p = 0.002).

Figure 4

SOX9 protein does not bind the chromosome axis of the LBCs. Immunostaining of P. waltl nuclear spreads using the anti-Cter-SOX9 antibody (green, Alexa 488 IgG) and counterstained with Hoechst (red). (A) Fluorescent micrograph (a) and its corresponding negative (b) showing a strong SOX9 immunostaining detected in the close vicinity of the chromosome axis over the thickest region of the lateral loops (B) Nuclear spreads from an oocyte at the end of stage VI. The SOX9 protein was not detected at the level of the chromosome axis where it was devoid of lateral loops (white arrows, merge panel). Scale bar for all micrographs: 10 μm.

Figure 5

Immunostaining of SOX9 after inhibition of transcription. Oocytes of P. waltl were treated or not with α-Amanitin or Actinomycin D (Act D). Phase contrast and corresponding fluorescent micrographs of nuclear spreads that were stained with the anti-Cter SOX9 antibody (green, Alexa 488 IgG) and counterstained with Hoechst (red) to show the chromosome axis. In contrast to the LBCs from the not-treated oocytes, those from the oocytes incubated with α-amanitin or Act D were devoid of their lateral loops and and the chromosome axis lacked SOX9 staining. Note that the spheres (arrowheads) known to be storage sites for transcription and post-transcription factors were strongly stained. The intensity of fluorescence in the different experiments was normalized with respect to their relevant controls with the secondary antibodies alone. Wide field Leica microscope. Scale bar for all micrographs: 15 μm.

Figure 6

Immunostaining of Pol II and SOX9 in P. waltl LBCs. Nuclear spreads were immunostained for SOX9 with the anti-Cter-SOX9 antibody (red, Alexa 568 IgG) and for Pol II with the mAbH14 (green, Alexa 488 IgM). Pol II immunostaining was continuous over the loop axis while that of SOX9 was concentrated on granules distributed above the loop axis. This double immunostaining pattern was particularly visible on the two lateral loops indicated by arrows in the boxed areas. The intensity of fluorescence in this experiment was normalized with respect to its relevant controls (i.e., secondary antibodies alone). The image was deconvoluted (see Figure S7 for the non deconvoluted image). Wide field Leica microscope. Scale bar: 10 μm.

Figure 7

Super-resolution images of a P. waltl lateral loop immunostained for CELF1 [mAb3B1 (red, Alexa 568 IgG)] and SOX9 [anti-Cter-SOX9 antibody (green, Alexa 488 IgG)]. The RNP matrix exhibited a pattern of granules immunostained for SOX9 and CELF1. The merged panel showed that the SOX9 and CELF1 granules did not colocalize. The intensity of fluorescence in the different experiments was normalized with respect to their relevant controls with the secondary antibodies alone. Structured Illumination Microscopy (SIM). The image in the boxed area corresponded to the phase contrast image obtained from Wide field Leica microscopy. Scale bar: 10 μm.

Figure 8

Localization of SOX9 to the matrix of lateral loops is RNA-dependent. Nuclear spreads of X. laevis were immunostained for SOX9 [anti-Cter-SOX9 antibody (green, Alexa 488 IgG)] and SR proteins [mAb1H4 (red, Alexa 568 IgG)]. (a–e’) LBC from untreated nuclear spreads. (f–j’) LBC from nuclear spread digested with RNase A before immunostaining. (a’–e’) Enlarged images of the regions marked by white boxes (a–e) show several lateral loops immunostained with SOX9 and SR antibodies. The arrowheads point to terminal granules. (f’–j’) enlarged images of the regions marked by white boxes (f–j) show one lateral loop not stained with SOX9 and SR antibodies after digestion with Rnase A. The intensity of fluorescence in the different experiments was normalized with respect to their relevant controls (the secondary antibodies only). Wide field Leica microscopy. Scale bar: 10 μm.