Germ plasm is found in germ-line cells of Xenopus and thought to include the determinant of primordial germ cells (PGCs). As mitochondria is abundant in germ plasm, vital staining of mitochondria was used to analyze the movement and function of germ plasm; however, its application was limited in early cleavage embryos. We made transgenic Xenopus, harboring enhanced green fluorescent protein (EGFP) fused to the mitochondria transport signal (Dria-line). Germ plasm with EGFP-labeled mitochondria was clearly distinguishable from the other cytoplasm, and retained mostly during one generation of germ-line cells in Dria-line females. Using the Dria-line, we show that germ plasm is reorganized from near the cell membrane to the perinuclear space at St. 9, dependent on the microtubule system.
Germ plasm is found in the germ-line cells of Caenorhabditis elegans, Drosophila and anuran amphibians (Ikenishi 1998). The only cells incorporating germ plasm differentiate into PGCs and finally sperm and eggs (Illmensee & Mahowald 1974; Ikenishi et al. 1986; Houston & King 2000; Tada et al. 2012). From the histological observation (Whitington & Dixon 1975; Akita & Wakahara 1985), Xenopus germ plasm is distributed patchily at the vegetal cortex of unfertilized eggs. Germ plasm is aggregated after fertilization, kept to the one pole of the cell membrane and parceled out into about four cells during cleavage stage. Aggregation of germ plasm and unequal division of blastomeres may be necessary to limit the number of cells bearing germ plasm. Shortly before the beginning of gastrulation, the germ plasm is reorganized to the perinuclear space from the cell cortex, accompanying equal cell division. This event may lead to determination of germ cells. As reorganization of germ plasm may represent processes of germ cell formation, live imaging would be required for further analysis of the dynamics of germ plasm.
In general, germ plasm is visualized by introducing a fusion construct comprising green fluorescent protein (GFP) cDNA and germ plasm-localizing genes: aubergine (Harris & Macdonald 2001) and vasa (Breitwieser et al. 1996; Sano et al. 2002) in Drosophila, plg-1 (Cheeks et al. 2004) in C. elegans, and vasa in medaka (Nagao et al. 2008). In Xenopus oocytes and embryos, germ plasm was visualized by expression of fusion mRNA, GFP and germ plasm-specific genes: Xpat, Germes, Dead end, and Xdazl, leading to inconvenient results coming for unkown reasons. The introduction of Xpat into oocytes induced the aggregation of mitochondria in somatic areas (Machado et al. 2005). Abnormal PGC formation and migration was induced by injection of Germes mRNA (Berekelya et al. 2007), and also of Dead end and Xdazl (K. Watanabe, unpubl. data, 2011). This method is not applicable to observe the normal behavior of germ plasm.
The germ plasm has abundant mitochondria, in contrast to the other cytoplasm of vegetal hemisphere. By vital staining of mitochondria, the movement of germ plasm was analyzed in early cleavage embryos (Savage & Danilchik 1993 and Venkatarama et al. 2010). However, the application of this method has been restricted to the early embryonic period.
In order to follow the behavior of germ plasm in development, it has been applied to Xenopus to label mitochondria by transgenic methods used in cultured cells (Horie et al. 2002). A transgene is EGFP fused to mitochondria transmembrane segment (TMS), which enables the expression of EGFP in the outer membrane of mitochondria. A selected transgenic Xenopus was nominated as Dria (mitochondria)-line. The EGFP protein in germ plasm was associated with germ plasm-specific molecules, and retained mostly during one generation of germ-line cells in Dria-line females. Using the Dria-line, we have confirmed the dynamics of germ plasm in living cells, as was suggested by histology (Whitington & Dixon 1975) and vital staining of germ plasm (the above papers). Furthermore, the reorganization of germ plasm just before gastrulation was analyzed in dissociated cells, and was found to depend on the microtubule system.
Materials and methods
Generation and screening of transgenic frogs
pRc/CMV EGFP-OMP25-transmembrane segment (TMS) enables EGFP expression on mitochondrial outermembrane, and is a gift from M. Sakaguchi (Horie et al. 2002). This construct was cut with BamHI and used for transgenesis according to the method described by Kroll & Amaya (1996) with modifications (Mizuno et al. 2005). Seventeen female transgenic F0 frogs were produced and crossed with wild-type males. F1 embryos of 14 out of 17 frogs expressed EGFP. Among them, four frogs were suitable for observation of germ plasm, which was clearly distinguished from the rest of the cytoplasm in their F1 blastomeres. We examined EGFP expression in the gonads and other organs of F1 froglets from the four F0 frogs. In one line (named the Dria-line), EGFP was expressed strongly in gonads, but weakly in the other tissues. In the other three lines, EGFP was expressed at a similar level in the gonads, kidney, skin and so on (data not shown). Oocytes of the Dria-line and also other lines seemed to express EGFP at a similar intensity. We decided that the Dria-line was especially suitable for observations of germ plasm.
Xenopus laevis eggs were fertilized in vitro as described previously (Kataoka et al. 2006). The fertilized eggs and embryos were kept in 0.1× Mark's modified ringer (MMR) at 18°C or 22°C. The larvae and adult frogs were reared at 23°C in water filtered through activated charcoal. The developmental stages were determined according to Nieuwkoop & Faber (1994). To produce F1 embryos, transgenic female frogs were crossed with wild-type males in most cases.
Culture of the dissociated cells
The culture of dissociated cells was performed as described previously (Ueno et al. 2006). Fertilization membranes of St. 8 embryos were manually removed with fine forceps in a modified low Ca2+ Stearn's solution (Ueno et al. 2006). The cells with germ plasm were picked up with a micropipet and transferred to a plastic dish coated with 1% agarose gel in the modified low Ca2+ Stearn's solution. The behavior of cells was observed under a stereomicroscope (Leica MZ16F). Confocal fluorescent images of the cells were acquired using a Zeiss LSM 510 upright confocal microscope. The nucleus is labeled with propidium iodide (PI).
Inhibition of the cytoskeleton; nocodazole (SIGMA), an inhibitor of microtubule polymerization, was dissolved in dimethyl sulfoxide (DMSO) at 30 mmol/L, and cytochalasin D (Cosmo Bio, Tokyo), an inhibitor of actin polymerization, was dissolved in methanol at 1 mmol/L. Nocodazole and cytochalasin D were used at final concentrations of 15 μmol/L and 1 μmol/L in the medium, respectively.
In situ hybridization
After fluorescence micrographs of living embryos and dissociated cells were taken, the samples were fixed in 25% methanol, 20% formalin at room temperature for 2 h and stored in 100% methanol at −20°C. In situ hybridization was performed using a digoxigenin (DIG)-labeled antisense RNA probe as described previously (Kataoka et al. 2005). The hybridization signal was detected using anti-digoxigenin-AP (Roche) and BM purple AP substrate (Roche). After detection of the signal, embryos were bleached with 1%H2O2, 5% formamide, and 0.5× sodium chloride/sodium citrate.
As for in situ hybridization of the dissociated cells, care was taken to identify the individual cell in the following way. A small number of cells was treated in a well. After the EGFP signal of each cell was photographed, in situ hybridization was performed. AP signal of each cell was photographed to align the orientation of the cells, similar to the previous EGFP image.
Immunocytochemistry of dissociated cells
For Xdazl staining, the dissociated cells were fixed in 2% PFA, 0.5 mol/L NaCl and 0.1 mol/L MOPS pH7.5 at 4°C overnight. For α- and γ-tubulin staining, the cells were fixed and preserved in −20°C methanol until used. Fixed cells were blocked for 2 h at room temperature in 10% goat serum in TPBS(−): 0.1%Triton X-100 dissolved in PBS (2.7 mmol/L KCl, 0.14 mol/L NaCl, 1.5 mmol/L KH2PO4, 8.1 mmol/L Na2HPO4). The samples were reacted with mouse monoclonal antibodies against Xdazl (Mita & Yamashita 2000), α-tubulin (calbiochem) and γ-tubulin (SIGMA) at 1:1000 for 1 day at 4°C in 10% goat serum/TPBS(−). They were washed three times with TPBS(−) for 30 min and reacted with Alexa546-conjugated anti-rabbit IgG (Molecular Probes) for 3 h at 22°C. Then they were washed three times with TPBS(−) for 30 min and stained with Hoechst 33342. After washing in PBS(−), they were transferred in 50% glycerol in PBS(−) and observed under a fluorescence microscope (BX60, Olympus).
Unfertilized eggs were fixed in 4% paraformaldehyde/PBS(−) at 4°C overnight, dehydrated in a graded ethanol-butanol series, and embedded in paraffin. Transverse 8 μm sections were de-paraffinized through a series of xylene, washed in PBS(−), blocked for 2 h at room temperature in 10% goat serum in TPBS(−) and, reacted with the mouse monoclonal antibody against glutamate oxaloacetate transaminase2 (GOT) (Kogo et al. 2011) and the rabbit polyclonal antibody against GFP (Clontech). After washing in TPBS(−), the sections were reacted with Alexa594-conjugated anti-mouse IgG and Alexa488-conjugated anti-rabbit IgG (Molecular Probes), respectively. After washing in TPBS(−), the specimens were mounted in 50% glycerol and observed under a fluorescent microscope BX60 (Olympus).
EGFP expression in germ plasm and germ-line cells
To trace the behavior of germ plasm in development, we generated transgenic Xenopus, Dria-line, harboring EGFP labeled mitochondria (see 'Materials and methods' for details). We examined EGFP expression during the life cycle, using F1 from crosses between Dria-line females and wild-type males. Enormous mitochondria exist in the animal hemisphere (Kogo et al. 2011), but EGFP expression of animal hemisphere was covered by the surface melanin pigments in St. 1 to St. 8 embryos. Whereas punctate EGFP signals were widely distributed around the vegetal pole of St.1 embryos and very weak signals were observed at most of the vegetal hemisphere (Fig. 1A). They were aggregated, forming strong fluorescent patches near the vegetal pole of St. 3, 6, 8 and 11 embryos (Fig. 1B–E). These strong EGFP signals displayed germ plasm, because they were colocalized with Xpat mRNA, a germ plasm-specific mRNA, at any stage examined (Fig. 1A′–E′). Other germ plasm-specific molecules were also colocalized with EGFP in germ plasm (cf. Fig. 6).
EGFP was surely expressed in mitochondria, because EGFP signals in the unfertilized eggs were colocalized with mitochondrial signals detected by mitochondria-specific anti-GOT antibody. In the animal hemisphere, mitochondria were abundant (Fig. 1F–F″). In the vegetal hemisphere, aggregated mitochondria showing germ plasm at the cell cortex was contrasted with a paucity of mitochondria in the inner vegetal cytoplasm (Fig. 1G–G″).
After St. 11, the endoderm including the germ plasm was enclosed by surface ectoderm. The EGFP became stronger around this stage (Fig. 1E). Finally germ plasm-containing cells (PGCs; primordial germ cells) resided inside of embryos. The trunks of neurula and tailbud embryos were dissociated into cells. The PGCs were distinguishable from numerous endodermal cells by EGFP expression: the former with strong signals and the latter with weak signals throughout the cell. PGCs in neurula expressed EGFP strongly in the perinuclear space and weakly throughout the cytoplasm (Fig. 1H, I, I′). Probable PGCs in dissociated cells judged from EGFP expression were positively stained with anti-Dazl antibody (Fig. 1J–J″). In the tailbud stage, The PGCs with strong signals showed the migrating and elongated forms (Fig. 1K, L, L′ arrows).
Furthermore, EGFP expression was examined in the later stage. Gonads with mesonephros were dissected from larvae and froglets. In St. 46 larvae, positive cells were found in gonads on the mesonephros (Fig. 2A, B, B′). EGFP expression gradually weakened in PGCs after St. 50 (data not shown). At St. 56, when sexual difference was distinguishable by pattern of pigment distribution, EGFP disappeared completely in all PGCs of both sexes (Fig. 2C, C′, F, F′). On the other hand, in somatic cells, EGFP expression was observed (see mesonephros behind the gonad). Then, at St. 60, the time of metamorphic climax, punctate signals were again expressed strongly in ovaries (Fig. 2D, D′ arrow), but not in testes (Fig. 2G, G′). The ovary formed several corrugations at St. 63, and contained a number of EGFP-positive cells (Fig. 2E, E′). EGFP signal was not detected in testes (Fig. 2H, H′). In females at 8 months post-fertilization, EGFP expression was strong throughout the ovary (Fig. 2I, I′). Pre-stage I and stage I oocytes were present and they contained strong EGFP signals in the precloud or mitochondrial cloud (Fig. 2J, J′). In male froglets, the EGFP level in testes was much lower than that in the surrounding somatic cells. After sexual maturation, EGFP was expressed weakly in the periphery of the testicular lobules (data not shown).
Maternal expression of transgene in germ-line
All mature oocytes laid by a Dria-line heterozygous female expressed the EGFP protein, translated from the maternal mRNA, and then all Dria-line eggs were fertilized by a wild-type sperms (Fig. 3A). Distinction of embryos with and without zygotic expression was difficult because of the maternal expression shortly after mid-blastula transition (MBT), but became clear at St. 27, when half of the embryos began to express strong EGFP protein in mitochondria-rich cells of epidermis (Brown et al. 1981). The other half gradually lost the EGFP signal from outside view because of a lack of zygotic expression (Fig. 3A, left panel). However, all individuals with and without the transgene similarly kept EGFP expression in germ plasm, and gradually lost EGFP until St. 56. In the case of individuals with transgene, EGFP expression in germ-line cells again began in oocytes of female gonads from St. 60 (Fig. 3A, right panel).
Mating of a Dria-line heterozygous male and a wild-type female produced fertilized eggs all without EGFP expression (Fig. 3B left panel). In somatic cells, half of the embryos with transgene began to express strong EGFP from St. 27 as mentioned above and the other half without transgene did not express EGFP. In germ-line cells of embryos with transgene, it was at St. 60 that the EGFP first began to express in oocytes of female gonads (Fig. 3B right panel, data not shown).
In germ-line, cytomegalovirus promoter works limitedly during oogenesis, and very weakly throughout spermatogenesis. In somatic line, it may work after MBT, though obvious expression of EGFP was observed from neurula.
Reorganization of germ plasm at gastrulation
Histological observation reported that germ plasm is reorganized to the perinuclear space from the cell cortex shortly before beginning of gastrulation (Whitington & Dixon 1975; Akita & Wakahara 1985). The reorganization of germ plasm at this stage is of particular interest, because it may be the beginning of close interaction between the germ plasm and nucleus, leading to the formation of PGCs. By live imaging of germ plasm, we first analyzed the participation of cytoskeleton to the reorganization of germ plasm. To determine when germ plasm is reorganized, about 10 embryos were fixed around the gastrula stage with a 2% paraformaldehyde solution and dissociated into cells. The PGCs were collected (2–9 cells per embryo) and the intracellular distribution of germ plasm was examined. Cells with little or dispersed germ plasm was ruled out, and only the cells clearly distinguished as PGCs were used. The germ plasm was condensed near the cell membrane (M-pattern) or expanded towards the center of the cell (C-pattern). The ratios of cells of M- or C-pattern to the total cells were indicated (Fig. 4A). At St. 8, 93.9% of cells were M-pattern and 6.1% were C-pattern, at St. 9, 55.0% were M-pattern and 45.0% were C-pattern, and at St. 10, 11.8% were M-pattern and 88.2% were C-pattern. The reorganization of germ plasm was shown to occur at St. 9, just before gastrulation as was proposed by histological observation (Whitington & Dixon 1975; Akita & Wakahara 1985). Next, we dissociated St. 8 living embryos into cells and observed the behavior of the germ plasm. The germ plasm was somewhat expanded near the cell membrane just after dissociation. It then gradually expanded towards the center of the cell, and surrounded the nucleus after reorganization (Fig. 4B). Even in the dissociated cells, reorganization of germ plasm occurred at a similar timing and frequency, as in the intact embryo (data not shown). It was found that the reorganization of germ plasm proceeded in most cases without cell division, and in a few cases accompanying cell division. Confocal images showed that the germ plasm was situated near the cell membrane, separately from nucleus before reorganization, and surrounded the nucleus after reorganization (Fig. 4C–C″, D–D″).
Requirement of microtubule polymerization for the reorganization of germ plasm
In general, the cytoskeleton plays an important role in the subcellular transport of organelles and mRNA. We examined whether the reorganization of germ plasm is driven by microfilament or microtubule system. The cells were dissociated from about 5–10 embryos at St.8 and only the PGCs with M-pattern were isolated, and treated with inhibitors. After 5 h treatment, the ratios of cells of M- or C-pattern to the total cells were indicated (Fig. 5A). Germ plasm was hardly reorganized in the presence of nocodazole (89.8%), an inhibitor of microtubule polymerization. In the cells treated with cytochalasin D, an inhibitor of actin polymerization, germ plasm was reorganized at the same level as the control (cytochalasin D: 83.3%, methanol: 91.1%). Additionally, to examine localization of germ plasm and microtubule components, we performed immunocytochemistry staining for α-tubulin and γ-tubulin. α-tubulin was colocalized with germ plasm throughout the reorganization process (Fig. 5B–B″, C–C″, D–D″). γ-tubulin resided in two separate areas before the reorganization: two centrioles near the nucleus, and germ plasm (Fig. 5E–E″). After the reorganization, almost all γ-tubulin resided in the perinuclear space (Fig. 5F–F″).
Then, reorganization of other germ plasm specific molecules was studied. Together with mitochondria (Fig. 6A–C), germ plasm-specific molecules, Xpat mRNA (Fig. 6A′), nanos1 mRNA (Fig. 6B′) and XDazl protein (Fig. 6C′), were also reorganized. In cells treated with nocodazole, these molecules remained at the cell membrane (Fig. 6D–F, D′–F′). Mitochondria and germ plasm-specific molecules were retained by microtubules and transported towards the nucleus by the extending microtubules.
Live imaging of germ plasm in the Dria-line
Live imaging using F1 females of the Dria-line revealed that EGFP-labeled germ plasm was distributed punctately at the vegetal cortex of St. 1 embryos, then aggregated and entered the embryo along the cleavage furrows as was shown histologically (Whitington & Dixon 1975). These strong signals were colocalized with Xpat mRNA, a germ plasm-specific molecule. The abundant mitochondria in germ plasm contrasted with the a little mitochondria in the vegetal hemisphere. This situation made the germ plasm visualize easily. It was difficult to observe germ plasm from outside of the gastrula after completion of epiboly. By dissociating blastomeres, we can easily identify PGCs by strong EGFP signals. In the Dria-line, germ plasm and PGCs behave as was revealed by histological observation in a wild-type embryo: reorganization of perinuclear space in neurula, motility of PGCs in the tailbud and settlement in genital ridges (Whitington & Dixon 1975). EGFP-labeled germ plasm can be traced during almost one generation of female germ-line cells. This Dlia-line lays healthy eggs and can be kept for generations.
Maternal expression of the transgene
The EGFP protein in germ plasm was produced and maintained during oogenesis, and maintained in a fertilized egg, in PGCs of the embryo and larva. The mRNA transcribed from the CMV promoter is only relieved during oogenesis. Maternal EGFP was likely stable in germ-line, compared with that in somatic cells. Kataoka et al. (2006) injected venus mRNA fused to 3′UTR of DEADSouth into the vegetal pole of St. 1 embryos, and found that Venus protein was stable in PGCs until St. 56. The metabolism in PGCs may be lower than that in somatic cells.
Reorganization of germ plasm and MBT
We have shown that the reorganization of germ plasm occurs at St. 9, just before gastrulation, by dissociating and culturing Dria-line blastomeres, as was reported in histological studies (Whitington & Dixon 1975; Akita & Wakahara 1985). In general, MBT occurs before St. 9, which includes lengthening of the cell cycle, activation of cell movement and initiation of zygotic transcription in somatic cells. In the case of germ-line cells, the cell cycle was prolonged at the time of the reorganization of germ plasm. Before the reorganization, it took 30 min for division, and asymmetrical distribution of germ plasm occurs. During the reorganization, it took at least 60 min (data not shown). Activation of cell movement and migration of germ plasm in the cell were accompanied with the reorganization. Accordingly, the reorganization of germ plasm might be one of the events included in MBT.
Reorganization is dependent on microtubules
γ-tubulin is a key component of microtubule-organizing centers (MTOCs) (Oakley & Oakley 1989; Stearns et al. 1991). γ-tubulin is contained in germ plasm, the so called mitochondrial cloud of stage I oocytes (Gard et al. 1995), and may elaborate the microtubule pavement when mitochondria and RNA in germ plasm move to the vegetal cortex of stage III and stage IV oocytes (Kloc & Etkin 1998). The vegetal cortex of stage VI oocytes contains γ-tubulin (Gard 1994) and mitochondria that are related to germ plasm (Kogo et al. 2011). Germ plasm-specific molecules localize at the vegetal apex (Mosquera et al. 1993; Houston et al. 1998; Hudson & Woodland 1998). After the activation of stage VI oocytes and fertilization, mitochondria and germ-plasm-specific molecules aggregate at the vegetal apex in association with the reorganization of microtubules (Ressom & Dixon 1988). Depletion of kinesin-like protein 1 (klp-1) by an antisense deoxyoligonucleotide inhibited the aggregation of germ plasm after fertilization, suggesting that the klp-1 motor protein carries components of germ plasm on microtubules (Robb et al. 1996). In the present study, we showed that α-tubulin and γ-tubulin accumulated in germ plasm in the vicinity of the cell membrane in St. 8 blastomeres, and that α-tubulin expanded toward the nucleus ahead of germ plasm during the reorganization at MBT. We further showed that the reorganization of germ plasm was blocked by an inhibitor of α-tubulin polymerization, but not by an inhibitor of actin polymerization. The reorganization of germ plasm in oogenesis, activation, fertilization and MBT was dependent on the polymerization of microtubules. Close interaction of the nucleus and surrounding germ plasm after MBT may establish the fate of PGCs. The Dria-line system using EGFP-labeled mitochondria would contribute to the further analysis of PGC formation.
This work was partly supported by the Ministry of Education, Science, Sports and Culture, Grant-in-Aid for Scientific Research (C) 23570259. We are grateful to Professor Masao Sakaguchi (Graduate School of Life Science, University of Hyogo) for pRc/CMV EGFP-OMP25 and to Professor Masakane Yamashita (Graduate School of science, Hokkaido University) for the anti-Xdazl monoclonal antibody. We also thank other members of our laboratory for their support.