Developmental Biology 364 (2012) 236–248

Contents lists available at SciVerse ScienceDirect

Developmental Biology

journal homepage: www.elsevier.com/developmentalbiology

Maternally localized germ plasm mRNAs and germ cell/stem cell formation in the cnidarian Clytia

Lucas Leclère a,⁎,1, Muriel Jager a,1, Carine Barreau b, Patrick Chang b, Hervé Le Guyader a, Michaël Manuel a, Evelyn Houliston b a Université Pierre et Marie Curie, Univ Paris 06 UMR 7138 CNRS MNHN IRD, Case 05, 7 quai St Bernard, 75005 Paris, France b Université Pierre et Marie Curie, Univ Paris 06 UMR 7009 CNRS, Observatoire Océanologique, 06230 Villefranche-sur-Mer, France article info abstract

Article history: The separation of the germ line from the soma is a classic concept in biology, and depending on spe- Received for publication 23 September 2011 cies is thought to involve fate determination either by maternally localized germ plasm (“preformation” or Revised 11 January 2012 “maternal inheritance”)orbyinductivesignaling(classicallytermed“epigenesis” or “zygotic induction”). Accepted 20 January 2012 The latter mechanism is generally considered to operate in non-bilaterian organisms such as cnidarians Available online 28 January 2012 and sponges, in which germ cell fate is determined at adult stages from multipotent stem cells. We have found in the hydrozoan cnidarian Clytia hemisphaerica that the multipotent “interstitial” cells (i-cells) in Keywords: Germ line larvae and adult medusae, from which germ cells derive, express a set of conserved germ cell markers: Multipotent stem cell Vasa, Nanos1, Piwi and PL10. In situ hybridization analyses unexpectedly revealed maternal mRNAs for Germ plasm all these genes highly concentrated in a germ plasm-like region at the egg animal pole and inherited by Preformation the i-cell lineage, strongly suggesting i-cell fate determination by inheritance of animal-localized factors. On the other hand, experimental tests showed that i-cells can form by epigenetic mechanisms in Clytia, since larvae derived from both animal and vegetal blastomeres separated during cleavage stages developed Evolution equivalent i-cell populations. Thus Clytia embryos appear to have maternal germ plasm inherited by i-cells but also the potential to form these cells by zygotic induction. Reassessment of available data indicates that maternally localized germ plasm molecular components were plausibly present in the common cnidarian/ bilaterian ancestor, but that their role may not have been strictly deterministic. © 2012 Elsevier Inc. All rights reserved.

Introduction of mRNAs and proteins (notably Piwi, Nanos, Vasa, PL10, Pumilio, Boule/Dazl and Bruno) involved in transposon silencing and mRNA In a wide variety of metazoan species, a distinctive maternal cyto- regulation (Ewen-Campen et al., 2010; Juliano et al., 2010a; Voronina plasmic region of the egg called pole plasm (in Drosophila) or germ et al., 2011). Germ plasm is detectable during oogenesis as an amor- plasm (Saffman and Lasko, 1999), is selectively inherited by the pri- phous substance termed nuage near the large oocyte nucleus (germinal mordial germ cells (PGC), founders of the germ line. In some cases, vesicle), and relocates to a restricted area of the cortex during oogenesis notably in Drosophila and in anuran amphibians such as Xenopus, or early development. It is then inherited by a subpopulation of blasto- germ plasm components have been shown experimentally to act in meres that give rise to the PGCs. the determination of the germ line. This mechanism of germ line Some animal species lack any distinguishable germ plasm during formation by inheritance of a maternal germ plasm is called “prefor- early embryonic stages, and their PGCs are specified by inductive mation” or “maternal inheritance”. signals (Extavour and Akam, 2003). This type of germ line formation Germ plasm can be recognized by various distinctive features (Eddy, is termed “epigenesis” or “zygotic induction” and is well documented 1975), including electron-dense granules composed of ribonucleo- in mammals and in urodele amphibians. In both cases, experimental protein complexes, variable association with dense concentrations manipulations can induce re-specification of cells from various embry- of mitochondria and nuclear pores and all or part of a conserved set onic regions to PGC fates, and there is no detectable mRNA or protein lo- calization for the germ plasm “markers”, or localization of electron- dense granules during early development (see Extavour and Akam, 2003). Bone morphogenetic proteins (BMPs) have been identified as ⁎ Corresponding author at: Sars International Centre for Marine Molecular Biology, primordial germ cell inducers in mouse embryos (Lawson et al., 1999; University of Bergen, Thormøhlensgate 55, N-5008, Bergen, Norway. E-mail address: [email protected] (L. Leclère). Ohinata et al., 2009; Ying et al., 2003), but there is no indication that 1 These authors contributed equally. they are involved in germ line specification in other species (reviewed

0012-1606/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2012.01.018 L. Leclère et al. / Developmental Biology 364 (2012) 236–248 237 in Ewen-Campen et al., 2010). Another “epigenetic” route to PGC forma- Material and methods tion is seen in sponges, hydrozoans and planarians, and involves their segregation from multipotent stem cell populations maintained in the Gene identification adult (Müller, 2006; Watanabe et al., 2009). Irrespective of the mode and timing of PGC specification, genes of cDNA sequences corresponding to the CheNanos1 and 2, ChePiwi, the Piwi, Vasa, nanos set have been consistently found expressed ei- ChePL10 and CheVasa genes were retrieved by BLAST searches on ther in the germ line or in multipotent PGC precursors. Furthermore, the Clytia hemisphaerica EST collection (publicly available on Gen- genes of this set have been shown to be essential for the maintenance Bank) sequenced by Genoscope (Evry, France) from C. hemisphaerica and differentiation of both PGC and multipotent stem cells (Ewen- mixed stage normalized cDNA libraries (see Houliston et al., 2010). Campen et al., 2010; Juliano et al., 2010a). Their expression, however, The PL10 sequence was incomplete and subsequently extended using is not exclusive to cells with germinal potential. For instance, Piwi, degenerate forward primers corresponding to amino-acid sequences Vasa and PL10 are expressed in some somatic stem cell types in MACAQT (PL10-1: 5′ ATGGCNTGYGCNCARAC 3′) and GSGKTAA (PL10- mammals and Drosophila (see Juliano et al., 2010a). In the cteno- 2: 5′ GGNWSNGGNAARACNGCNGC 3′) and specific reverse primers phore Pleurobrachia, expression of Piwi, Vasa and PL10 genes occurs (PL10-1rev: 5′ ATCCAACGCGACCAACAGCC 3′ and PL10-2rev: 5′ GCTAA- in both the germ line and in a variety of non-germ line stem cells CATCTGAATTTCC 3′). Two rounds of nested PCR were performed using, (Alié et al., 2011). as template, 1 μl of diluted cDNA extracted from a Clytia cDNA library. Expression of the germ plasm/germ line/stem cell genes described GenBank accession numbers: EU199802 (ChePiwi), JQ397273 (CheVasa), above has been used to trace the embryological origin of the germ JQ397274 (CheNanos1), JQ397275 (CheNanos2) and JQ397276 (ChePL10). line, and to infer its mechanism of specification. The curious scattered distribution of species thus inferred to use “preformation” versus Phylogenetic analyses “epigenesis” across the animal phylogeny has stimulated much de- bate as to which mechanism is evolutionarily the oldest. A survey Cnidarian and bilaterian sequences were retrieved from GenBank based heavily on such gene expression data suggested that “epigene- or at www.compagen.org (Hemmrich and Bosch, 2008). Sequences sis” was more widespread and probably ancestral (Extavour, 2007; were aligned using CLUSTALW in the BioEdit package (Hall, 1999)and Extavour and Akam, 2003). the alignment corrected manually. Conserved blocks were extracted Among the non-bilaterian metazoan lineages, hydrozoan cnidarians to perform phylogenetic analyses, carried out using the Maximum- are the group for which the origin of germ cells is best understood. Likelihood (ML) method using the PhyML program (Guindon and Cnidarian are divided in two clades, anthozoans and medusozoans Gascuel, 2003) with the JTT model of amino-acid substitutions (Jones which include hydrozoans (Collins et al., 2006). Medusozoans are et al., 1994). A BioNJ tree was used as the input tree to generate the characterized by the presence of a medusa, in addition to the ML tree. Among-site variation was estimated using a discrete approxi- stage. In anthozoans such as , the embryonic origin of mation to the gamma distribution with 8 rate categories. The gamma the germ cells and stem cells is unclear. No germ cell-generating shape parameter and the proportion of invariant sites were optimized pluripotent stem cells equivalent to hydrozoan interstitial cells (i-cells) during the ML search. Branch support was tested with bootstrapping have been detected (Technau and Steele, 2011), while reports of (100 replicates). the localization of maternal Nanos mRNAs are contradictory (Extavour et al., 2005; Torras and Gonzalez-Crespo, 2005). In hydrozoans, the and embryo manipulation life cycle typically comprises three phases: the planula larva, the benthic, vegetatively-propagating polyp and the pelagic, sexual me- We used Clytia hemisphaerica medusae cultured in Villefranche- dusa. The i-cell population, present throughout the adult life, gener- sur-Mer from established laboratory colonies as described previously ates both gamete precursors and various somatic cell types, namely (Chevalier et al., 2006). 8-cell stage or blastula stage embryos were neuro-sensory cells (including nematocytes) and secretory gland cut using fine tungsten needles on 2% agarose-coated Petri dishes. cells (Watanabe et al., 2009). The i-cells originate during gastrula- Embryo fragments were cultured in Millipore filtered natural or arti- tion and are first detectable in the central, endodermal region. They ficial seawater containing antibiotics on agarose coated Petri dishes. are later found predominantly in the ectoderm, or between the ecto- Although C. hemisphaerica embryos exhibit variable morphologies dermal and endodermal epithelia in the polyp and medusa. Studies during early development, we were able to use the “peanut” shape in various hydrozoan species (Podocoryne, Hydractinia and ) most common among C. hemisphaerica blastulae as an indicator of have shown that i-cells in the planula larva, polyp and medusa ex- the animal–vegetal axis (see Video S1). press genes considered to be germ line markers in bilaterian species (Piwi, Nanos, Vasa, PL10: Mochizuki et al., 2000, 2001; Rebscher et al., In situ hybridization 2008; Seipel et al., 2004) and contain dense cytoplasmic granules similar to those considered characteristic of germ plasm (Noda and Single and double in situ hybridizations were performed using DIG- Kanai, 1977 in Hydra). The impressive capacity of hydrozoan isolat- or fluorescein-labeled antisense RNA probes as described in (Denker ed fragments from both early and late stage embryos to regulate et al., 2008) but with two modifications. (i) Color was developed with normal development (e.g., Freeman, 1981), has encouraged the as- NBT/BCIP (Roche, Indianapolis, USA) for simple in situ hybridization, sumption that i-cells have an epigenetic origin. Recent reports of or NBT/BCIP and Fast RedTR-naphthol reagent (Sigma) for double in maternally-localized Vasa protein in the hydrozoan Hydractinia situ hybridizations. (ii) The concentration of each probe in the hy- (Rebscher et al., 2008), however, has raised doubts about this issue. bridization buffer was adapted to obtain the best results (low back- We have addressed the origin of i-cells in the experimental model ground and intense signal): Piwi (80 ng/μl), Nanos1 (20 ng/μl), Nanos2 Clytia hemisphaerica (Houliston et al., 2010). In situ hybridization an- (20 ng/μl), PL10 (2 ng/μl), Vasa (8 ng/μl) with 1 μl used for hybridiza- alyses of five germ line marker genes (Piwi, Nanos1, Nanos2, PL10 and tion in a final volume of 1 ml. Vasa) along with embryo bisection and q-PCR analyses provided evi- dence that maternally localized germ plasm co-exists with an epige- Transmission electron microscopy netic type mechanism of i-cell specification during Clytia embryonic development. Our findings have prompted us to reconsider the rela- Embryos and gonads were pre-fixed at room temperature (RT) for tionship between germ plasm and germ line in metazoans, as well 10 min in solution A (3% glutaraldehyde, 0.3 M NaCl, 0.05% OsO4, as between preformation and epigenesis. 0.1 M sodium cacodylate pH 7.3), then rinsed for 5 min in solution R 238 L. Leclère et al. / Developmental Biology 364 (2012) 236–248

(0.3 M NaCl, 0.2 M sodium cacodylate pH 7.3). They were fixed for 2 h Random Hexamer Primer and Transcriptor Reverse Transcriptase at RT in solution A without OsO4, rinsed in solution R for 5 min and (Roche Applied Science, Indianapolis, USA). q-PCRs were run in trip- post-fixed for 1 h on ice in 1% OsO4, 1.5% K-ferricyanide, 0.3 M NaCl, licate or quadruplicate and EF-1alpha used as the reference control 2.5% NaHCO3, pH 7.2 (protocol modified after Eisenman and Alfert, gene. Each PCR contained 0.8 μl cDNA, 10 μl SYBR Green I Master 1982 and Sun et al., 2007). Samples were rinsed in H2O and dehy- Mix (Roche Applied Science), and 200 nM of each gene-specific prim- drated for 15 min each, with an ethanol series (50, 70, 90, 95%) fol- er, in a 20 μl final volume. q-PCR reactions were run in 96-well plates, lowed by 4 incubations in 100% ethanol. They were then embedded in a LightCycler 480 Instrument (Roche Applied Science). Negative in Spurr resin (Agar). Semi-thin sections were stained with 0.5% controls (RNA without reverse transcriptase) were performed for Methylene blue in 1% Borax. Thin sections were counter-stained every primer pair in each PCR plate, to ensure the absence of genomic with saturated aqueous uranyl acetate (15 min) and Reynolds lead DNA amplification. Sequences of forward and reverse primers designed citrate (15 min). for each gene: EF-1alpha-F 5′ TGCTGTTGTCCCAATCTCTG 3′; EF-1alpha- R5′ AAGACGGAGTGGTTTGGATG 3′;piwi-F5′ GGTCACGACCCAGA- Quantitative RT-PCR CAGAAT 3′;piwi-R5′ GGAATGAGCGAAAAGACGAG 3′;wnt3-F5′ ATCATGGCAGGTGGAAACTC 3′;wnt3-R5′ CCCCATTTCCAACCTTCTTC Total RNA from individual cells or embryos was extracted using 3′; vasa-F 5′ GCTCGTGCACTTGTCAAAAC 3′; vasa-R 5′ ACCCCAGC- RNAqueous-Micro kit according to the manufacturer's instructions CATCATTGTTAC 3′; PL10-F 5′ ACTGGTTTGTCCACCTCGTT 3′; PL10-R (Ambion, Warrington, UK). First-strand cDNA was synthesized using 5′ CACCGCCATATCTGCTCTTT 3′;nanos1-F5′ AATGAACCCTGGACCTTTCC

Fig. 1. Expression of stem cell marker genes in Clytia medusae. A: In situ hybridization staining of whole medusae (in A1—arrow: one of the four gonads, arrowhead: one of the eight tentacle bulbs, asterisk: the manubrium). Insert in A4: staining in a statocyst (delineated by a dotted line). B, C: Expression in differentiating spermatozoids and oocytes within isolated gonads (or non-isolated in C4). In CI, C2 and C5 the gonad structure had been opened out and flattened so that the early oogenesis stages positioned proximally form a peripheral ring. D, E: Tentacle bulbs oriented with the proximal side at the top and the tentacle on the bottom. The row E shows double in situ hybridizations in tentacle bulbs with stem cell marker genes (blue: B) and Minicollagen 3-4a (red: R) probes. Co-staining makes a purple color (P). Insert in E4: higher magnification of CheNanos2–Minicollagen co-expressing cells. Scale bars: A1–A5: 100 μm; B1–E5: 25 μm; insert in A4 and E4: 10 μm. L. Leclère et al. / Developmental Biology 364 (2012) 236–248 239

3′;nanos1-R5′ TCACTGTCTTTGAGCGTGTG 3′;nanos2-F5′ GCCATCT- (Figs. 1E1–E3, E5), consistent with data from Hydra (Mochizuki et al., CAACCACAAAACC 3′;nanos2-R5′ AATCGGGCAAGTGTAAGCAC 3′. 2000, 2001)andPodocoryne (Seipel et al., 2004). CheNanos2 was expressed in germ cells like the other genes, but otherwise differed in expression. In the tentacle bulb, CheNanos2 Imaging mRNA was detected in a central band showing complete overlap with expression of minicollagen (Fig. 1E4), indicating that this gene All images of in situ-stained specimens were acquired on an Olympus is expressed in differentiating nematoblasts but not in i-cells. It was BX61 microscope using a Q-imaging Camera with Image Pro plus soft- also detected in the statocyst basal epithelium, near the bell margin ware (Mediacybernetics, Bethesda, MD). TEM images were acquired on (insert in Fig. 1A4). In young but not in mature medusae, CheNanos2 a Hitachi H-600 at 100 kV. Timelapse recordings were made on a Zeiss expression was also detected in the endoderm of the radial and circular Axiovert microscope with a motorized stage and camera driven by gastrovascular canals and of the manubrium (Figs. 1A4, S3). Metamorph software. Localized maternal mRNAs in a germ plasm-like domain, inherited by Results i-cells

“Germ line” gene expression in jellyfish germ cells and somatic stem cells The maternal mRNAs for all five genes studied were found to be concentrated around the nucleus in small and large growing oocytes We identified orthologues of five potential germ line/stem cell (Figs. 2A–E). In this region, transmission electron microscope (TEM) genes from our C. hemisphaerica EST collection (Nanos (CheNanos1, sections revealed amorphous “nuage” associated with mitochondria CheNanos2), PL10 (ChePL10), Vasa (CheVasa), Piwi (ChePiwi)—see and nuclear pores (Figs. 2F, G), highly reminiscent of germ plasm Figs. S1 and S2 for gene orthology analyses). described in other species (Eddy, 1975). The maternal mRNAs for In situ hybridization of Clytia medusae using antisense probes for all five genes investigated also showed a marked asymmetric distri- all five genes investigated in this study showed expression in the bution in spawned eggs, being highly concentrated in a restricted zone of proliferating germ cell progenitors in the gonad (Amiel and region immediately adjacent to the female pronucleus at the animal Houliston, 2009). In addition, except for CheNanos2, they all showed pole (Figs. 3A1–A5). In Hydractinia, perinuclear localization of Vasa expression in “somatic” stem cells at the base of the tentacle bulbs protein has been described in the corresponding region (Rebscher (Fig. 1), specialized swellings that continuously supply nematocytes et al., 2008). (among other cell types) to the tentacles (Denker et al., 2008). In During cleavage stages, transcripts of all the 5 studied genes except this study we will consider both of these two spatially-separate CheVasa remained strongly detectable in the cortical region of blasto- stem cell populations as i-cells since they show equivalent gene ex- meres inherited from the animal pole of the egg (Figs. 3B–D). In most pression, but whether they are functionally interchangeable remains cases, the mRNAs were distributed in two patches on either side of to be verified experimentally. As in Podocoryne (Seipel et al., 2004; the animal pole of cleavage and blastula stage embryos (see Fig. 3, Torras et al., 2004), Hydractinia (Rebscher et al., 2008) and Hydra lines C and D, but note the single patch in Figs. 3D1, D4). The split of (Mochizuki et al., 2000, 2001), expression of all these genes was the maternal mRNA into two patches reflects the division by the more easily detectable in the gonad than at other sites. In both male first unipolar cleavage furrow cutting through the animal pole, and (Figs. 1B1–B5) and female (Figs. 1C1–C5) gonads, the in situ signal for frequent subsequent physical separation of the animal sides of the all these genes was most intense in a proximal strip and distal rim, cor- two first blastomeres as development proceeds (see Video S1). responding to the sites of presumptive stem cells and early oocyte dif- In early gastrula-stage embryos, ChePiwi, CheNanos1, CheNanos2 and ferentiation stages in females. Within tentacle bulbs, ChePiwi, ChePL10 mRNA were detected in a cluster of small cells at the oral pole, CheNanos1, CheVasa and ChePL10 expression was detected both in the site of cell ingression (Fig. 3E1–E4). We hypothesize that zygotic the stem cell zone of the proximal bulb area (see Denker et al., transcription of these genes starts at the late blastula–early gastrula 2008, Figs. 1D1–D3, D5) and during the earliest step of nematogenesis, stage, as has been suggested for other embryonically expressed genes as indicated by limited overlap with expression of minicollagen (e.g., Momose et al., 2008). At the end of gastrulation (approximately

Fig. 2. Perinuclear concentration of “germ plasm” in growing oocytes. A–E: Characteristic perinuclear distribution of the five mRNAs in very early oocyte stages (cyt: cytoplasm; nu: nucleus). F, G: TEM sections of Clytia hemisphaerica growing oocytes. Red arrowheads indicate perinuclear “nuage” material, similar to that described in the germ plasm of many bilaterian animals. er: endoplasmic reticulum, nm: nuclear membrane, pm: cell membrane, n: nucleus, nu: nucleolus, mit: mitochondria. Scale bars A–E, G: 5 μm, F: 1 μm. 240 L. Leclère et al. / Developmental Biology 364 (2012) 236–248

Fig. 3. Distribution of stem cell marker mRNAs during embryonic development. Continuity of ChePiwi, ChePL10, CheNanos1 and CheNanos2 mRNAs from animal cortex of the egg to i- cells. CheVasa is more broadly expressed during cleaving stages and becomes restricted to i-cell during gastrulation and early planula formation. In all panels the position of the animal/oral pole is marked by an asterisk and corresponds to the gastrulation initiation site (asterisk in gastrulae E1–E5). Insert in B5: higher magnification of perinuclear CheVasa mRNA detection in an 8-cell stage embryo. Scale bars for all panels: 50 μm.

20–24 h post-fertilization) they were predominantly found in the distribution of these cells changed progressively, such that after oral half of the endodermal region of the newly-formed planula 2daysofdevelopmentChePiwi, CheNanos1, CheNanos2 and ChePL10 larva (see Fig. 3F1–F4 and Fig. 4). Over the following 24 h the positive cells were found dispersed throughout the endoderm of L. Leclère et al. / Developmental Biology 364 (2012) 236–248 241

Fig. 4. ChePiwi in situ hybridization during post-blastula embryonic development. A–C: 3 successive stages of gastrulation. D, E: 1 day planulae, F: 2 day planula, G: 3 day planula. H: Schematic representation of embryos shown in panels A to G (dark gray: ectoderm, light gray: endoderm, white: blastocoel, blue: ChePiwi positive cells). Embryos get longer and thinner during gastrulation and planula formation. Flattening of the planula during micro-slide preparation explains why they look bigger than early gastrula in the panels D–G; the size of the planula has been corrected in the drawings shown in H. I: High magnification of Piwi-positive putative i-cells in the endodermal region of a 3 day old planula (ec: ec- toderm, en: endoderm, n: nucleus). Scale bars: A–G: 50 μm; I: 10 μm.

Fig. 5. Piwi expression in embryos derived from blastula halves. In situ hybridization staining for ChePiwi on uncut control embryos (A) fixed in parallel with embryos derived from lateral (B), animal or vegetal (C) halves of mid-blastula stage embryos and fixed 5–30 min, 6 h or 20 h after cutting. Panels C4 and C5 show two different in situ hybridization pat- terns obtained with embryo deriving from the vegetal half at t=0–30 min. In all pictures the animal/oral pole is positioned on top. Proportions of embryos with or without clear Piwi-positive cells following fixation at successive times after cutting (n=number of embryos scored) are indicated on top left in each panel. Scale bars for all panels: 50 μm. 242 L. Leclère et al. / Developmental Biology 364 (2012) 236–248 the planula larva. In 2 and 3 day planulae Nanos2 expression was also (5–6 h post fertilization), the animal pole of such embryos is marked detected in the aboral pole ectoderm (Fig. 3F4). by a deep cleft between two major lobes (e.g., see Fig. 3D2), which In 3 day planulae, ChePiwi, CheNanos1, CheNanos2 and ChePL10 from time-lapse movies of development can be deduced to derive positive cells in the now-differentiated endoderm layer were identifi- from the animal side of the first unipolar cleavage furrow (Video able as i-cells by their round shape, characteristic disposition, and S1). In peanut-shaped embryos, in which two patches of ChePiwi, high nuclear–cytoplasmic ratio (Figs. 3 F1–F4, 4I). Interstitial cells have CheNanos1, CheNanos2 and ChePL10 mRNA-rich cells are situated been described in the endodermal region of mature C. hemisphaerica towards the tips of the major two lobes, we attempted bisection planulae in histological studies (Bodo and Bouillon, 1968). In Hydra, perpendicular to the cleft to separate animal and vegetal fragments i-cell derivatives have been shown to include nematocytes, nerve or, for comparison, along the cleft to generate lateral halves each cells, gland cells, germ cells and differentiating stages of these four cell containing one lobe and therefore one patch of Piwi/Nanos1/Nanos2/ types (Watanabe et al., 2009). No trace of capsule, dense granules or PL10-RNA positive cells (see diagrams in Figs. 5B and C). Our attempts neurite-like structures was associated with the cells expressing the at producing animal–vegetal separation were partially successful, four genes in the Clytia larvae. We thus conclude that these four genes with 10/17 (58.8%) embryos deriving from blastula “vegetal” halves are expressed in multipotent i-cells and possibly also early stages of dif- fixed immediately after cutting (within 30 min) lacking Piwi-mRNA ferentiation of derivative cell types, but not in mature nematoblasts, aggregates, while 100% of embryos derived from “animal” (n=17) gland or nerve cells. and “lateral” (n= 50) halves showed Piwi staining. The 7/17 Piwi- The expression pattern for CheVasa during embryonic develop- positive “vegetal” halves obtained in our experiments most probably ment was slightly different from the other genes studied. In the egg, resulted from inaccurate cutting, but could theoretically also reflect CheVasa mRNA was detected, albeit weakly, in the same cortical re- rapid re-expression of Piwi in some vegetal fragments. Despite the gion at the animal pole as the other mRNAs (Fig. 3A5). In embryos successful elimination of Piwi-mRNA aggregates in nearly 60% of and larvae, however, cytoplasmic transcripts appeared distributed in cases, nearly all gastrula (6 h after cutting; 20/22; 91%) and planula a broad animal–vegetal gradient, with diffuse accumulations around (20 h after cutting; 5/5; 100%) stage embryos derived from the “veg- nuclei in early cleavage stages (Fig. 3B5). In gastrulae and planulae, etal” halves, showed clear populations of Piwi positive cells a population of cells strongly expressing CheVasa was detected with (Figs. 5C6 and C7), in a very similar pattern as uncut controls or the same general distribution as the putative i-cells expressing the lateral halves (Figs. 5A2, A3, B2 and B3). Furthermore, following other four genes (Figs. 3E5, F5). culture until the mature planula stage, all larvae derived from both animal and vegetal halves contained morphologically distinguish- Experimental demonstration of i-cell formation by epigenesis able nematocytes and gland cells. These blastula bisection experiments indicate that a functional To test whether i-cells can form in the absence of the Piwi/Nanos1/ i-cell lineage can develop in the absence of maternally-derived Nanos2/PL10/Vasa mRNA-rich “germ plasm”, embryo bisection exper- “germ plasm”, but are open to the criticism that there was signifi- iments were performed at the 8-cell and blastula stages. Develop- cant contamination of the “vegetal” fragments by germ plasm. To ment of the i-cell population in the resultant half embryos was ensure that none of the animal germ plasm material contaminated monitored by in situ hybridization, using ChePiwi as a marker. the vegetal fragments, we thus separated 8-cell stage embryos into We first attempted to separate the animal and vegetal halves at single blastomeres. At this stage the germ plasm mRNA patches are the blastula stage by taking advantage of the characteristic peanut restricted to the 4 animal blastomeres, and so if germ plasm determines shape adopted by many cleaving embryos. At the blastula stage cell fate no i-cells should develop from the vegetal blastomeres (Fig. 6).

Fig. 6. Piwi expression in embryos derived from isolated 8-cell stage blastomeres. A, B: In situ hybridization staining for ChePiwi on uncut control embryos (A) fixed in parallel with mini-embryos (B) derived from isolated blastomeres of 8 cell stage embryos at different times of development: 5–20 min, 1 h, 6.5 h, 20 h and 50 h after cutting. In all panels the animal/oral pole is on top. Proportion of embryos with or without clear Piwi-positive cells following fixation at successive times after cutting (n=number of embryos scored) are indicated on top left in each panel. ChePiwi in situ hybridization of 2 day old planula (A5) comes from an independent experiment. Scale bars for all panels: 50 μm. L. Leclère et al. / Developmental Biology 364 (2012) 236–248 243

Fig. 7. Widespread distribution of “germ plasm” mRNAs in early embryos. Quantitative RT-PCR detection of CheWnt3, CheNanos1, CheNanos2, ChePL10, ChePiwi and CheVasa mRNAs in (A) animal and vegetal halves from three individual mid-blastula stage embryos (numbered 1 to 3 on the X axis) and (B) 4 isolated blastomeres from animal or vegetal halves of four 8 cell stage embryos (numbered 1 to 4), processed immediately after cutting. mRNA levels in each half are expressed as a percentage of the total quantity in the embryo. The animal cortical localized mRNA Wnt3 was quantified in parallel for comparison. q-PCR was performed in triplicate or quadruplicate and results normalized with respect to the level of EF-1alpha mRNA.

As expected, Piwi mRNA aggregates were detectable in 35/66 (53%) of Discussion blastomeres fixed within 20 min of isolation (Fig. 6B1). Note that the isolates were mixed, because in the absence of clear morphological In this study we have provided evidence that maternally localized polarity markers it is hard to generate large populations of uniquely “germ plasm” mRNAs and a regulative “zygotic induction” (“epige- animal or vegetal blastomeres. Correspondingly, at the blastula stage, netic”) mechanism of interstitial stem cell lineage specification may 6 h 30 min after blastomere isolation, strongly Piwi positive cells were co-exist during embryogenesis in the hydrozoan Clytia hemisphaerica. detected in only 28/60 (47%) of the resultant mini-embryos (Fig. 6B3) The maternally localized mRNAs are not inherited by a dedicated indicating that there had been no reformation of germ plasm during germ line, but appear to segregate into precursors of multipotent pre-blastula development. In planula larvae, however, fixed 20 h (1- stem cells (i-cells). We suggest that Clytia germ plasm does not day planula) or 50 h (2-day planula) after fertilization, 100% of the have a strictly deterministic function, but may favor the generation embryos had correctly formed epithelial endoderm and ectoderm of i-cell precursors from the animal region of the egg during embry- layers and nearly all displayed Piwi-positive cells (73/78 and 65/68 onic development, thus coordinating their formation with that of respectively—Figs. 6B4 and B5). Furthermore, mature nematocytes the endoderm. Regulative mechanisms can promote i-cell formation were visible at the aboral pole in 66% (43/65) of the 2-day planulae when germ plasm is missing and could also account for position- with Piwi-positive cells in the endodermal region. These results dependent i-cell formation during normal embryogenesis. Distinct demonstrate that embryo fragments lacking detectable mRNA-rich mechanisms based on signaling from surrounding tissues at a much “germ plasm” are able to re-generate Piwi-expressing i-cells by the later life cycle stage segregate definitive PGCs from the i-cell popula- time of gastrulation, and thereby to develop a functional interstitial tion. Our findings have prompted us to reconsider the relationship be- cell lineage in the endoderm. tween germ plasm and germ line, as well as between preformation/ maternal inheritance mechanisms and two types of epigenetic/zygotic induction mechanisms, one operating during embryogenesis and the A pool of non-localized maternal germ cell/stem cell mRNAs other in adult life.

The observed development of Piwi-positive i-cells in vegetal em- bryo fragments could theoretically involve either the mobilization of Maternal germ plasm in Clytia? non-localized maternal mRNAs or new transcription of germ cell/ stem cell genes. As a first step to distinguishing these possibilities We have uncovered a distinct region in the animal cytoplasm of we quantified mRNA levels by reverse transcription PCR (Q-RT- Clytia eggs characterized by marked local concentrations of Vasa, PCR). Mid-blastula stage embryos were bisected into animal and veg- Piwi, Nanos and PL10 mRNAs. This region of the cytoplasm overlaps etal halves and subject directly to PCR to quantify levels of ChePiwi, with a domain of cortical Wnt3 mRNA localization, which extends fur- CheVasa, ChePL10, CheNanos1 and CheNanos2 mRNAs. The experiment ther away from the animal pole (Amiel and Houliston, 2009; Momose shown in Fig. 7A involved independent measurement of three et al., 2008), and also with the localization of CheSox1 and CheSox13 matched pairs of animal and vegetal halves. CheNanos1 and CheNa- mRNAs (Jager et al., 2011). The maternal perinuclear Piwi, Nanos nos2 mRNAs showed a clear animal bias in their distribution, equiva- and PL10 mRNA aggregates are inherited by a distinct cell population lent to that of CheWnt3, whose maternal mRNAs has been shown to located at the animal pole through cleavage stages. During gastrula- be highly localized at the animal cortex of eggs and early embryos tion, expression of Piwi, Nanos1 and PL10, along with Vasa, continues (Momose et al., 2008), while others showed levels in the vegetal zygotically in a sub-population of cells corresponding to i-cells (and half that were only modestly lower than in animal halves (ChePL10, possibly some of their undifferentiated derivatives) in the planula ChePiwi) or indistinguishable (CheVasa). Equivalent results were larvae. This maternal mRNA localization and inheritance profile is obtained following bisection of four 8-cell stage embryos along the highly suggestive of a “maternal inheritance” mechanism generating third cleavage plane. The average proportion of mRNA in the vegetal i-cells. Although the i-cells do not constitute a dedicated germ line, half (4 blastomeres) was approximately 30% for CheNanos1, 40% for we propose to retain the term “germ plasm”,todefine a characteristic CheNanos2, ChePL10 and ChePiwi and 50% for CheVasa (Fig. 7B). cytoplasmic domain inherited by a “germ track”,whichincludes Prior to the blastula stage, Clytia embryos thus contain significant both multipotent somatic/germinal stem cells and dedicated germ vegetal pools of many “germ plasm” mRNAs, despite the locally cell precursors. The concept of the germ track was originally elabo- high concentration around the animal pole revealed by in situ rated by August Weismann (1893) in relation to his theory of the hybridization. continuity of the “germ plasm”, originally developed from 244 L. Leclère et al. / Developmental Biology 364 (2012) 236–248 L. Leclère et al. / Developmental Biology 364 (2012) 236–248 245 observations of germ cell segregation in hydrozoans (Berrill and Liu, “Zygotic induction” mechanisms for i-cell formation 1948; Weismann, 1883). It referred to the genealogy of cells contain- ing the germ plasm from the egg to the germ cells, with or without Our q-PCR measurements indicate that in Clytia, the region of i-cell early segregation of a proper germ line, i.e. of a cell lineage that formation is not restricted spatially by the distribution of “germ plasm” does not produce any somatic cell. It is important to realize that mRNAs. Although they are highly concentrated in the animal pole region modern use of the term germ plasm, as referring to a cytoplasmic these mRNAs are also present across all regions of the early embryo structure, is completely different to Weismann's germ plasm, to a greater or lesser extent. The pools of non-localized maternal RNAs which in fact equates with nuclear genetic material (reviewed in clearly cannot play a classical determinant role otherwise all cells would Lankenau, 2008). However, since the continuity of the (cytoplasmic) become i-cells. A very similar situation exists for Drosophila germ plasm germ plasm appears to occur in species with or without early segre- components, with the majority of nanos and oskar mRNAs dispersed gation of the germ line, the concept of “germ track” could find a new throughout the embryo cytoplasm (96% and 82% respectively) despite relevance. striking mRNA localization to posterior pole plasm revealed by in situ hy- It is important to emphasize that the proposition that the Vasa/ bridization (Bergsten and Gavis, 1999). Nanos/Piwi/PL10 mRNA-rich “germ plasm” in Clytia acts as a maternal Given the presence of a distinct domain with germ plasm charac- determinant, and so directs the fate of the cells that inherit it, is teristics at the egg animal pole, our experimental demonstration that currently based only on circumstantial evidence, and has two im- a regulative mechanism can generate i-cells from vegetal regions of portant caveats. Firstly, current evidence does not show definitive- cleavage or blastula stage embryos was unexpected. Although initial- ly that the i-cell precursors actually inherit the maternally localized ly uncomfortable at a conceptual level, co-existence of germ plasm- germ plasm mRNA, although they do form in the appropriate posi- based and “zygotic induction” mechanisms for interstitial stem cell tion in the embryo. Confirmation of direct inheritance during the lineage specification during embryogenesis can be easily reconciled blastula–gastrula transition would require the maternal mRNAs or at a mechanistic level (as already suggested by Extavour, 2007). Start- the cells that inherit them at the blastula stage to be tracked in vivo, ing from a situation where a conserved set of stem cell genes such as possibilities that are not at the moment technically feasible. The succes- Piwi, Nanos and Vasa is expressed in the oocyte, it is easy to imagine sive segregation of these localized mRNA during cleavage divisions, that the aggregation and association of their maternal mRNAs in means that lineage tracing by dye injection at an earlier stage would one part of the egg could favor rapid interstitial stem cell determina- be uninformative. tion in the cells that inherit them, while signaling mechanisms acting Secondly, it should be stressed that localization does not imply after the onset of zygotic gene transcription could contribute to reg- function. Even if they are inherited by the i-cells, it is possible that ulating the size of the cell population and/or to regionally restricting the localized “germ plasm” mRNAs have no role in determining their formation and/or proliferation, or to facilitate their restoration their fate, or indeed no significant function in development, as following stress or injury. A similar situation has been described in seems to be the case for a set of mRNAs similarly closely associated the ascidian Ciona intestinalis where a maternal germ plasm seems with female pronucleus in the beetle Tribolium (Peel and Averof, to contain determinants of the germ line but removal of the PGCs 2010). It is also possible that they could participate in other at larval stage can be compensated by formation of new ones developmental processes, for example, in embryonic axis formation as from multipotent stem cells (Takamura et al., 2002). seen with Drosophila Nanos. Testing the function of maternally The term “zygotic induction” comprises very diverse modes of germ localized mRNAs in Clytia is currently problematic because their line formation and is defined in opposition to the “maternal inheri- proximity to the nucleus precludes transplantation and irradiation tance” mechanism. This later mode can clearly be defined based on approaches. Testing the roles of individual genes by using classical observations and more recent molecular analyses: (i) pres- morpholino antisense oligonucleotides to block translation could ence of a distinct, restricted cytoplasmic region generated during be undertaken (e.g. Momose et al., 2008), but would affect both oogenesis, or at least before the first division; (ii) this specialized maternally and zygotically transcribed mRNAs and so would not cytoplasm is restricted to a distinct subset of cells during early devel- distinguish between roles in maternal inheritance and regulative opment, and (iii) the cells that take up this cytoplasm differentiate mechanisms. Specific experimental approaches to disrupt germ into the primordial germ cells, while those that are not associated plasm formation during oogenesis, and/or transgenesis approaches with the germ plasm, at least initially, take on a somatic fate. Those to specifically track germ plasm molecular components need to be organisms where all three criteria are met can be said to use mater- developed. nal inheritance, while those that do not, including Clytia,require

Fig. 8. Phylogenetic distribution of maternal localized germ plasm molecular components in the egg and type of PGC segregation in Metazoa. The distribution of germ plasm RNAs and/or proteins (in red) is represented at 1-cell stage, cleavage and gastrula embryonic stages for one species per taxonomic group for which molecular data are available. For groups containing both species with maternal localized germ plasm molecular components and species without, a species with maternal localization is presented. At larval/juvenile stage, the presence and distribution of the germ line (in yellow) and/or multipotent stem cell line giving rise to the germ line (in blue) are represented. Asterisks indicate the an- imal/anterior pole for bilaterians and the animal/oral pole for cnidarians and ctenophores. Metazoan phylogeny presented is derived from Philippe et al. (2011). For character cod- ing, only groups for which molecular data are available were considered: Porifera (Funayama et al., 2010; Müller, 2006—species represented: Suberites domuncula), Ctenophores (Alié et al., 2011—sp.: Pleurobrachia pileus), Anthozoa (Extavour et al., 2005; Torras and Gonzalez-Crespo, 2005—sp.: Nemastostella vectensis/staining shown for early embryonic stage: Vasa, PL10, Nanos2), Hydrozoa (this study; Rebscher et al., 2008; Torras et al., 2004; Seipel et al., 2004; Mochizuki et al., 2000, 2001—sp.: Clytia hemisphaerica/st.: Piwi, PL10, Nanos), Annelida (Agee et al., 2006; Dill and Seaver, 2008; Giani et al., 2011; Kang et al., 2002; Pilon and Weisblat, 1997; Rebscher et al., 2007; Sugio et al., 2008; Tadokoro et al., 2006—sp.: Platynereis dumerilii/st.: Vasa), Mollusca (Fabioux et al., 2004; Rabinowitz et al., 2008; Swartz et al., 2008; Kranz et al., 2010—sp.: Crassostrea gigas/st.: Vasa), Roti- fera (Smith et al., 2010—sp.: Brachionus plicatilis/st.: Vasa), Nematoda (Subramaniam and Seydoux, 1999; Salinas et al., 2007—sp.: Caenorhabditis elegans/st.: VBH-1=PL10), Arthro- poda (Hay et al., 1990; Chang et al., 2002; 2009; Dearden et al., 2003; Extavour, 2005; Sagawa et al., 2005; Dearden, 2006; Schröder, 2006; Nakkrasae and Damrongphol, 2007; Mito et al., 2008; Özhan-Kizil et al., 2009—sp.: Drosophila melanogaster/st.: Vasa, Nanos), Chaetognatha (Carré et al., 2002—sp.: Sagitta setosa/st.: Vasa), Echinodermata (Voronina et al., 2008; Juliano et al., 2006; Juliano et al., 2010b; Yajima and Wessel, 2011—sp.: Strongylocentrotus purpuratus/st.: Vasa), Cephalochordata (Wu et al., 2011—sp.: Branchiostoma flor- idae/st.: Vasa, Nanos), Urochordata (Takamura et al., 2002; Shirae-Kurabayashi et al., 2006; Sunanaga et al., 2006; Sunanaga et al., 2008; Brown et al., 2009, Review in Kawamura et al., 2011—sp.: Ciona intestinalis/st.: Vasa), Teleostei (Knaut et al., 2002; Raz, 2003; Saito et al., 2004; Herpin et al., 2007; Aoki et al., 2008—sp.: Danio rerio/st.: Vasa), Amphibia Anura (Machado et al., 2005; Sekizaki et al., 2004—sp.: Xenopus laevis/st: Vasa), Amphibia Caudata (Tamori et al., 2004; Bachvarova et al., 2004—sp.: Ambystoma mexicanum/st.: dazl), Mammalia (review for mice in Hayashi et al., 2007; Lee et al., 2005—sp.: Mus musculus/st.: Nanos3), Aves (Tsunekawa et al., 2000—sp.: Gallus gallus/st.: Vasa). We did not consider molecular studies in Acoelomorpha (De Mulder et al., 2009), Platyhelminthes (Handberg-Thorsager and Saló, 2007; Pfister et al., 2008; Sato et al., 2006), turtles (Bachvarova et al., 2009), Dipnoi (Johnson et al., 2003) and Chondrostei (Johnson et al., 2011) since these publications do not provide expression patternsinmatureeggsandcleavagestageembryos. Note that while the Rotifera are considered in this figure to lack localized maternal germ plasm on the basis of the only molecular study available (Smith et al., 2010), this con- clusion is provisional since pre-molecular studies suggest that localized maternal germ plasm is present at least in some species of this group (see Extavour and Akam, 2003). 246 L. Leclère et al. / Developmental Biology 364 (2012) 236–248 additional input for the induction of the germ cells, and thus use the present in non-vertebrate genomes such as Ciona and Bran- “zygotic induction” mode. chiostoma that display “maternal inheritance” (Shirae-Kurabayashi The “zygotic induction” modes of germ line formation can be di- et al., 2006; Wu et al., 2011). vided into different types. One type, classically observed in mammals Given the various arguments in both directions and the lack of and in urodele amphibians, involves determination of PGC fate during functional testing of the role of maternal germ plasm mRNAs in embryogenesis, with no role for maternal germ plasm. A second type most metazoan phyla, it is not possible for the moment to deduce is found in species like Clytia. A maternal germ plasm has a facultative whether the ancestral function of maternal germ plasm was to pro- or essential role in multipotent stem-cell lineage determination, but vide determinants of the germ track. We suggest that both localized distinct intercellular signaling based mechanisms act at a much later maternal germ plasm and “zygotic induction” of the germ track dur- life cycle stage to segregate definitive PGCs from the multipotent ing embryonic development might have co-existed in the ancestor stem cell population. A similar two-step model of PGC determination of the Eumetazoa and that some animal lineages have subsequently was proposed by Rebscher et al. (2007) based on observations in the favored or lost one or the other during evolution. This would offer polychaete Platynereis in which multipotent stem cells seem to orig- an explanation of their presence in a diverse and phylogenetically inate by inheritance of maternal determinants, and definitive germ dispersed set of animals. cell fate determination occurs secondarily, through an epigenetic Maternal germ plasm RNAs in Clytia, as well as early zygotic ex- signal, from the multipotent stem cells. pression in Nematostella (Extavour et al., 2005), are localized on the same side as mRNAs that direct gastrulation and endoderm formation The evolutionary origin of maternal germ plasm (Martindale, 2005; Momose et al., 2008). The relationship between germ plasm and gastrulation is thus the same in cnidarians and in As demonstrated in the survey presented in Fig. 8, maternally bilaterians, although the fate map is reversed with respect to egg localized germ plasm as assessed mainly from molecular data, is animal–vegetal polarity (Martindale, 2005). A common association very widespread in the Metazoa. Germ plasm ultrastructure and between the germ plasm and the gastrulation site might represent its position in the egg at the future site of gastrulation are an ancient feature of animal development, facilitating association of very similar between distant animal groups. Germ plasm con- germ cell precursors with developing endodermal or mesodermal tains variable combinations of conserved gene products present derivatives, and/or reflecting ancestral participation of germ plasm as mRNAs or proteins or both depending on the species, such components in embryo patterning. The case of Nanos is interesting that no single molecular component is a universal feature. Even in this context as this mRNA has an additional function in patterning in species lacking maternal localization, germ plasm components the posterior pole in various protostomes (e.g. Drosophila—Lehmann can localize very early during embryogenesis, for instance in mi- and Nusslein-Volhard, 1991, grasshopper—Lall et al., 2003, mollusk— cromeres at the 16-cell stage in echinoderms (Voronina et al., Rabinowitz et al., 2008, and leech—Agee et al., 2006). The role of 2008; Yajima and Wessel, 2011) or in the 4d cell lineage for most Nanos in axial patterning may have evolved in the protostome line- species of annelids and molluscs (e.g. Dill and Seaver, 2008; Kranz age since no such function has been described in a deuterostome, et al., 2010; Swartz et al., 2008). Overall, rare are the groups without and remains to be tested in non-bilaterians. localization of germ plasm components in a germ track prior to gastru- Considering the phylogenetic distribution of species harboring lation. Previous phylogenetic surveys, in which detection of maternal early germ line segregation versus post-embryonic PGC formation germ plasm was equated with preformation, and evidence for late from a multipotent stem cell lineage such as hydrozoan i-cells PGC formation equated with epigenesis, concluded that maternal germ (Fig. 8), it is not clear which type of germ track is ancestral (i.e. plasm/preformation probably derived from epigenesis multiple times one-step or two-step PGC segregation). Several recent reviews and in evolution (Extavour, 2007; Extavour and Akam, 2003). articles (e.g. Juliano and Wessel, 2010; Rebscher et al., 2007)favor Given the close relationship of the Bilateria and Cnidaria an ancestral multipotent stem cell system capable of giving rise to (Philippe et al., 2009; Philippe et al., 2011), and the presence of lo- both germ line and somatic cells and still present in hydrozoans calized germ plasm in both Clytia and in nearly all the major bila- and planarians. Caution is required, however, since this scenario re- terian clades (Fig. 8), the hypothesis that maternally localized lies heavily on the absence of early embryonic PGC segregation in germ plasm was present in the last common cnidarian–bilaterian many animal groups for which data on early embryonic stages are (=eumetazoan) ancestor remains quite plausible. This “ancestral very scattered and contradictory, like Platyhelminthes, Acoelomor- germ plasm” scenario implies that maternal germ plasm was lost pha or Anthozoa (De Mulder et al., 2009; Extavour et al., 2005; in the urodele amphibians and mammalian lineages as well as in Pfister et al., 2008). To solve this question, it would be particularly the anthozoans. In support of this scenario, “basally branching” informative to know the timing of germ line segregation in early such as urochordates or cephalochordates and proto- branching metazoans such as anthozoans, acoels and ctenophores. stomes such as chaetognaths seem to have a maternal germ plasm and “maternal inheritance” mode of germ line specification (Carré et Acknowledgments al., 2002; Shirae-Kurabayashi et al., 2006; Wu et al., 2011). On the other hand, in insects, maternal germ plasm and early setting We thank our research colleagues, Elsa Denker, Lisbeth C. Olsen aside of germ cells are restricted to the monophyletic Holometabola, and Fabian Rentzsch for useful suggestions, and two anonymous re- and are associated with the presence and function of the novel gene viewers for comments that improved the quality of the manuscript. oskar, which is an Holometabola specificgene(Lynch et al., 2011). This work was supported by a grant from the GIS “Institut de la Species that have lost oskar also have lost maternal inheritance. This Génomique Marine”–ANR “programme blanc” NT_NV_52 Genocni- molecular and phylogenetic evidence indicates that maternal germ daire and by the “Agence Nationale de la Recherche” grant ANR-09- plasm seen in insects such as Drosophila is a derived state within the BLAN-0236-01 DiploDevo. EST sequencing was performed by the insects. In the chordate lineage, maternal inheritance is associated with Consortium National de Recherche en Génomique at the Genoscope another novel gene, bucky-ball, which is similarly restricted to the (Evry, France). vertebrate lineage (Bontems et al., 2009). This might also indicate that maternal germ plasm in this clade is an evolutionary novelty Appendix A. Supplementary data associated with the invention of a new gene. However this gene is also present in mammalian genomes that have a clear epigenetic Supplementary data to this article can be found online at doi:10. mode of PGC segregation (Extavour and Akam, 2003) and is not 1016/j.ydbio.2012.01.018. L. Leclère et al. / Developmental Biology 364 (2012) 236–248 247

References Guindon, S., Gascuel, O., 2003. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst. Biol. 52, 696–704. Agee, S.J., Lyons, D.C., Weisblat, D.A., 2006. Maternal expression of a NANOS homolog is Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis required for early development of the leech Helobdella robusta. Dev. Biol. 298, program for Windows 95/98/NT. Nucleic Acids Symp. Ser. 41, 95–98. 1–11. Handberg-Thorsager, M., Saló, E., 2007. The planarian nanos-like gene Smednos is Alié, A., Leclère, L., Jager, M., Dayraud, C., Chang, P., Le Guyader, H., Quéinnec, E., expressed in germline and eye precursor cells during development and regener- Manuel, M., 2011. Somatic stem cells express Piwi and Vasa genes in an adult ation. Dev. Genes Evol. 217, 403–411. ctenophore: ancient association of “germline genes” with stemness. Dev. Biol. Hay, B., Jan, L.Y., Jan, Y.N., 1990. Localization of vasa, a component of Drosophila polar 350, 183–197. granules, in maternal-effect mutants that alter embryonic anteroposterior polarity. Amiel, A., Houliston, E., 2009. Three distinct RNA localization mechanisms contribute to Development 109, 425–433. oocyte polarity establishment in the cnidarian Clytia hemisphaerica. Dev. Biol. 327, Hayashi, K., de Sousa, Chuva, Lopes, S.M., Surani, M.A., 2007. Germ cell specification in 191–203. mice. Science 316, 394–396. Aoki, Y., Nagao, I., Saito, D., Ebe, Y., Kinjo, M., Tanaka, M., 2008. Temporal and spatial lo- Hemmrich, G., Bosch, T.C., 2008. Compagen, a comparative genomics platform for early calization of three germline-specific proteins in medaka. Dev. Dyn. 237, 800–807. branching metazoan animals, reveals early origins of genes regulating stem-cell Bachvarova, R.F., Masi, T., Drum, M., Parker, N., Mason, K., Patient, R., Johnson, A.D., differentiation. Bioessays 30, 1010–1018. 2004. Gene expression in the axolotl germ line: Axdazl, Axvh, Axoct-4, and Axkit. Houliston, E., Momose, T., Manuel, M., 2010. Clytia hemisphaerica: a jellyfish cousin Dev. Dyn. 231, 871–880. joins the laboratory. Trends Genet. 26, 159–167. Bachvarova, R.F., Crother, B.I., Manova, K., Chatfield, J., Shoemaker, C.M., Crews, D.P., Jager, M., Quéinnec, E., Le Guyader, H., Manuel, M., 2011. Multiple Sox genes are Johnson, A.D., 2009. Expression of Dazl and Vasa in turtle embryos and ovaries: expressed in stem cells or in differentiating neurosensory cells in the hydrozoan evidence for inductive specification of germ cells. Evol. Dev. 11, 525–534. Clytia hemisphaerica. Evodevo 2, 12. Bergsten, S.E., Gavis, E.R., 1999. Role for mRNA localization in translational activation Johnson, A.D., Drum, M., Bachvarova, R.F., Masi, T., White, M.E., Crother, B.I., 2003. Evo- but not spatial restriction of nanos RNA. Development 126, 659–669. lution of predetermined germ cells in vertebrate embryos: implications for macro- Berrill, N.J., Liu, C.K., 1948. Germplasm, Weismann, and hydrozoa. Q. Rev. Biol. 23, evolution. Evol. Dev. 5, 414–431. 124–132. Johnson, A.D., Richardson, E., Bachvarova, R.F., Crother, B.I., 2011. Evolution of the germ Bodo, F., Bouillon, J., 1968. Étude histologique du développement embryonnaire de line–soma relationship in vertebrate embryos. Reproduction 141, 291–300. quelques hydroméduses de Roscoff: Phialidium hemisphaericum (L.), Obelia sp. Jones, D.T., Taylor, W.R., Thornton, J.M., 1994. A model recognition approach to the prediction Péron et Lesueur, Sarsia eximia (Allman), Podocoryne carnea (Sars), Gonionemus of all-helical membrane protein structure and topology. Biochemistry 33, 3038–3049. vertens Agassiz. Cah. Biol. Mar. 9, 69–104. Juliano, C., Wessel, G., 2010. Developmental biology. Versatile germline genes. Science Bontems, F., Stein, A., Marlow, F., Lyautey, J., Gupta, T., Mullins, M.C., Dosch, R., 2009. 329 (5992), 640–641. Bucky ball organizes germ plasm assembly in zebrafish. Curr. Biol. 19, 414–422. Juliano, C.E., Voronina, E., Stack, C., Aldrich, M., Cameron, A.R., Wessel, G.M., 2006. Germ Brown, F.D., Tiozzo, S., Roux, M.M., Ishizuka, K., Swalla, B.J., De Tomaso, A.W., 2009. Early line determinants are not localized early in sea urchin development, but do accu- lineage specification of long-lived germline precursors in the colonial ascidian mulate in the small micromere lineage. Dev. Biol. 300, 406–415. Botryllus schlosseri. Development 136, 3485–3494. Juliano,C.E.,Swartz,S.Z.,Wessel,G.M.,2010a.Aconservedgermlinemultipotency Carré, D., Djediat, C., Sardet, C., 2002. Formation of a large Vasa-positive germ granule and its program. Development 137, 4113–4126. inheritance by germ cells in the enigmatic Chaetognaths. Development 129, 661–670. Juliano, C.E., Yajima, M., Wessel, G.M., 2010b. Nanos functions to maintain the fate of Chang, C.C., Dearden, P., Akam, M., 2002. Germ line development in the grasshopper the small micromere lineage in the sea urchin embryo. Dev. Biol. 337, 220–232. Schistocerca gregaria: vasa as a marker. Dev. Biol. 252, 100–118. Kang, D., Pilon, M., Weisblat, D.A., 2002. Maternal and zygotic expression of a nanos- Chang, C.C., Huang, T.Y., Cook, C.E., Lin, G.W., Shih, C.L., Chen, R.P., 2009. Developmental class gene in the leech Helobdella robusta: primordial germ cells arise from seg- expression of Apnanos during oogenesis and embryogenesis in the parthenogenetic mental mesoderm. Dev. Biol. 245, 28–41. pea aphid Acyrthosiphon pisum.Int.J.Dev.Biol.53,169–176. Kawamura, K., Tiozzo, S., Manni, L., Sunanaga, T., Burighel, P., De Tomaso, A.W., 2011. Chevalier, S., Martin, A., Leclère, L., Amiel, A., Houliston, E., 2006. Polarised expression Germline cell formation and gonad regeneration in solitary and colonial ascidians. of FoxB and FoxQ2 genes during development of the hydrozoan Clytia hemisphaerica. Dev. Dyn. 240, 299–308. Dev. Genes Evol. 216, 709–720. Knaut, H., Steinbeisser, H., Schwarz, H., Nüsslein-Volhard, C., 2002. An evolutionary Collins, A.G., Schuchert, P., Marques, A.C., Jankowski, T., Medina, M., Schierwater, B., conserved region in the vasa 3′UTR targets RNA translation to the germ cells in 2006. Medusozoan phylogeny and character evolution clarified by new large and the zebrafish. Curr. Biol. 12, 454–466. small subunit rDNA data and an assessment of the utility of phylogenetic mixture Kranz, A.M., Tollenaere, A., Norris, B.J., Degnan, B.M., Degnan, S.M., 2010. Identifying the models. Syst. Biol. 55, 97–115. germline in an equally cleaving mollusc: Vasa and Nanos expression during embry- Dearden, P.K., 2006. Germ cell development in the honeybee (Apis mellifera); vasa and onic and larval development of the vetigastropod Haliotis asinina. J. Exp. Zool. B nanos expression. BMC Dev. Biol. 6, 6. Mol. Dev. Evol. 314, 267–279. Dearden, P., Grbic, M., Donly, C., 2003. Vasa expression and germ-cell specification in Lall, S., Ludwig, M.Z., Patel, N.H., 2003. Nanos plays a conserved role in axial patterning the spider mite Tetranychus urticae. Dev. Genes Evol. 212, 599–603. outside of the Diptera. Curr. Biol. 13, 224–229. De Mulder, K., Kuales, G., Pfister, D., Willems, M., Egger, B., Salvenmoser, W., Thaler, M., Lankenau, D.H., 2008. The legacy of the germ line — maintaining sex and life in metazoans: Gorny, A.K., Hrouda, M., Borgonie, G., Ladurner, P., 2009. Characterization of the cognitive roots of the concept of hierarchical selection. In: Egel, R., Lankenau, D.-H. stem cell system of the acoel Isodiametra pulchra. BMC Dev. Biol. 9, 69. (Eds.), Genomic Dynamics and Stability. : Recombination and Meiosis, 3. Springer, Denker, E., Manuel, M., Leclère, L., Le Guyader, H., Rabet, N., 2008. Ordered progression Berlin, Heidelberg, pp. 289–339. of nematogenesis from stem cells through differentiation stages in the tentacle Lawson, K.A., Dunn, N., Roelen, B.A., Zeinstra, L.M., Davis, A.M., Wright, C.V.E., Korving, bulb of Clytia hemisphaerica (Hydrozoa, Cnidaria). Dev. Biol. 315, 99–113. J.P.W.F.M., Hogan, B.L.M., 1999. Bmp4 is required for the generation of primordial Dill, K.K., Seaver, E.C., 2008. Vasa and nanos are coexpressed in somatic and germ line germ cells in the mouse embryo. Genes Dev. 13, 424. tissue from early embryonic cleavage stages through adulthood in the polychaete Lee, G.S., Kim, H.S., Lee, S.H., Kang, M.S., Kim, D.Y., Lee, C.K., Kang, S.K., Lee, B.C., Hwang, Capitella sp. I. Dev. Genes Evol. 218, 453–463. W.S., 2005. Characterization of pig vasa homolog gene and specific expression in Eddy, E.M., 1975. Germ plasm and the differentiation of the germ cell line. Int. Rev. germ cell lineage. Mol. Reprod. Dev. 72, 320–328. Cytol. 43, 229–280. Lehmann, R., Nusslein-Volhard, C., 1991. The maternal gene nanos has a central role in Eisenman, E.A., Alfert, M., 1982. A new fixation procedure for preserving the ultra- posterior pattern formation of the Drosophila embryo. Development 112, 679–691. structure of marine invertebrate tissues. J. Microsc. 125, 117–120. Lynch, J.A., Ozüak, O., Khila, A., Abouheif, E., Desplan, C., Roth, S., 2011. The phylogenetic Ewen-Campen, B., Schwager, E.E., Extavour, C.G.M., 2010. The molecular machinery of origin of oskar coincided with the origin of maternally provisioned germ plasm and germ line specification. Mol. Reprod. Dev. 77, 3–18. pole cells at the base of the Holometabola. PLoS Genet. 7, e1002029. Extavour, C.G.M., 2005. The fate of isolated blastomeres with respect to germ cell for- Machado, R.J., Moore, W., Hames, R., Houliston, E., Chang, P., King, M.L., Woodland, H.R., mation in the amphipod crustacean Parhyale hawaiensis. Dev. Biol. 277, 387–402. 2005. Xenopus Xpat protein is a major component of germ plasm and may func- Extavour, C.G., 2007. Evolution of the bilaterian germ line: lineage origin and modula- tion in its organisation and positioning. Dev. Biol. 287, 289–300. tion of specification mechanisms. Integr. Comp. Biol. 47, 770–785. Martindale, M.Q., 2005. The evolution of metazoan axial properties. Nat. Rev. Genet. 6, Extavour, C.G., Akam, M., 2003. Mechanisms of germ cell specification across the meta- 917–927. zoans: epigenesis and preformation. Development 130, 5869–5884. Mito, T., Nakamura, T., Sarashina, I., Chang, C.C., Ogawa, S., Ohuchi, H., Noji, S., 2008. Dy- Extavour, C.G., Pang, K., Matus, D.Q., Martindale, M.Q., 2005. vasa and nanos expression namic expression patterns of vasa during embryogenesis in the cricket Gryllus patterns in a sea anemone and the evolution of bilaterian germ cell specification bimaculatus. Dev. Genes Evol. 218, 381–387. mechanisms. Evol. Dev. 7, 201–215. Mochizuki, K., Sano, H., Kobayashi, S., Nishimiya-Fujisawa, C., Fujisawa, T., 2000. Ex- Fabioux, C., Huvet, A., Lelong, C., Robert, R., Pouvreau, S., Daniel, J.Y., Minguant, C., Le pression and evolutionary conservation of nanos-related genes in Hydra. Dev. Pennec, M., 2004. Oyster vasa-like gene as a marker of the germline cell develop- Genes Evol. 210, 591–602. ment in Crassostrea gigas. Biochem. Biophys. Res. Commun. 320, 592–598. Mochizuki, K., Nishimiya-Fujisawa, C., Fujisawa, T., 2001. Universal occurrence of the Freeman, G., 1981. The role of polarity in the development of the hydrozoan planula vasa-related genes among metazoans and their germline expression in Hydra. larva. Dev. Genes Evol. 190, 168–184. Dev. Genes Evol. 211, 299–308. Funayama, N., Nakatsukasa, M., Mohri, K., Masuda, Y., Agataa, K., 2010. Piwi expression Momose, T., Derelle, R., Houliston, E., 2008. A maternally localised Wnt ligand required for in archeocytes and choanocytes in demosponges: insights into the stem cell system axial patterning in the cnidarian Clytia hemisphaerica. Development 135, 2105–2113. in demosponges. Evol. Dev. 12, 275–287. Müller, W.E., 2006. The stem cell concept in sponges (Porifera): metazoan traits. Semin. Giani Jr., V.C., Emi, Y., Michael, B.J., Seaver, E.C., 2011. Somatic and germline expression Cell Dev. Biol. 17, 481–491. of piwi during development and regeneration in the marine polychaete annelid Nakkrasae, L.I., Damrongphol, P., 2007. A vasa-like gene in the giant freshwater prawn, Capitella teleta. Evodevo 2, 10. Macrobrachium rosenbergii. Mol. Reprod. Dev. 74, 835–842. 248 L. Leclère et al. / Developmental Biology 364 (2012) 236–248

Noda, K., Kanai, C., 1977. An ultrastructural observation on Pelmatohydra robusta at Smith, J.M., Cridge, A.G., Dearden, P.K., 2010. Germ cell specification and ovary struc- sexual and asexual stages, with a special reference to “Germinal plasm”. J. Ultra- ture in the rotifer Brachionus plicatilis. Evodevo 1, 5. struct. Res. 61, 284–294. Subramaniam, K., Seydoux, G., 1999. nos-1 and nos-2, two genes related to Drosophila Ohinata, Y., Ohta, H., Shigeta, M., Yamanaka, K., Wakayama, T., Saitou, M., 2009. A sig- nanos, regulate primordial germ cell development and survival in Caenorhabditis naling principle for the specification of the germ cell lineage in mice. Cell 137, elegans. Development 126, 4861–4871. 571–584. Sugio, M., Takeuchi, K., Kutsuna, J., Tadokoro, R., Takahashi, Y., Yoshida-Noro, C., Özhan-Kizil, G., Havemann, J., Gerberding, M., 2009. Germ cells in the crustacean Par- Tochinai, S., 2008. Exploration of embryonic origins of germline stem cells and hyale hawaiensis depend on Vasa protein for their maintenance but not for their neoblasts in Enchytraeus japonensis (Oligochaeta, Annelida). Gene Expr. Patterns formation. Dev. Biol. 327, 230–239. 8, 227–236. Peel, A.D., Averof, M., 2010. Early asymmetries in maternal transcript distribution asso- Sun, M.G., Williams, J., Munoz-Pinedo, C., Perkins, G.A., Brown, J.M., Ellisman, M.H., ciated with a cortical microtubule network and a polar body in the beetle Tribolium Douglas, R., Green, D.R., Frey, G.T., 2007. Correlated three-dimensional light and castaneum. Dev. Dyn. 239, 2875–2887. electron microscopy reveals transformation of mitochondria during apoptosis. Pfister, D., De Mulder, K., Hartenstein, V., Kuales, G., Borgonie, G., Marx, F., Morris, J., Nat. Cell Biol. 9, 1057–1065. Ladurner, P., 2008. Flatworm stem cells and the germ line: developmental and evo- Sunanaga, T., Saito, Y., Kawamura, K., 2006. Postembryonic epigenesis of Vasa-positive lutionary implications of macvasa expression in Macrostomum lignano. Dev. Biol. germ cells from aggregated hemoblasts in the colonial ascidian, Botryllus primi- 319, 146–159. . Dev. Growth Differ. 48, 87–100. Philippe, H., Derelle, R., Lopez, P., Pick, K., Borchiellini, C., Boury-Esnault, N., Vacelet, J., Sunanaga, T., Satoh, M., Kawamura, K., 2008. The role of Nanos homologue in gameto- Renard, E., Houliston, E., Quéinnec, E., Da Silva, C., Wincker, P., Le Guyader, H., Leys, genesis and blastogenesis with special reference to male germ cell formation in the S., Jackson, D.J., Schreiber, F., Erpenbeck, D., Morgenstern, B., Wörheide, G., Manuel, colonial ascidian, Botryllus primigenus. Dev. Biol. 324, 31–40. M., 2009. Phylogenomics revives traditional views on deep animal relationships. Swartz, S.Z., Chan, X.Y., Lambert, J.D., 2008. Localization of Vasa mRNA during early Curr. Biol. 19, 706–712. cleavage of the snail Ilyanassa. Dev. Genes Evol. 218, 107–113. Philippe, H., Brinkmann, H., Lavrov, D.V., Littlewood, D.T., Manuel, M., Wörheide, G., Tadokoro, R., Sugio, M., Kutsuna, J., Tochinai, S., Takahashi, Y., 2006. Early segregation of Baurain, D., 2011. Resolving difficult phylogenetic questions: why more sequences germ and somatic lineages during gonadal regeneration in the annelid Enchytraeus are not enough. PLoS Biol. 9, e1000602. japonensis. Curr. Biol. 16, 1012–1017. Pilon, M., Weisblat, D.A., 1997. A nanos homolog in leech. Development 124, 1771–1780. Takamura, K., Fujimura, M., Yamaguchi, Y., 2002. Primordial germ cells originate from the Rabinowitz, J.S., Chan, X.Y., Kingsley, E.P., Duan, Y., Lambert, J.D., 2008. Nanos is re- endodermal strand cells in the ascidian Ciona intestinalis. Dev. Genes Evol. 212, 11–18. quired in somatic blast cell lineages in the posterior of a mollusk embryo. Curr. Tamori, Y., Iwai, T., Mita, K., Wakahara, M., 2004. Spatio-temporal expression of a DAZ- Biol. 18, 331–336. like gene in the Japanese newt Cynops pyrrhogaster that has no germ plasm. Dev. Raz, E., 2003. Primordial germ-cell development: the zebrafish perspective. Nat. Rev. Genes Evol. 214, 615–627. Genet. 4, 690–700. Technau, U., Steele, R.E., 2011. Evolutionary crossroads in developmental biology: Rebscher, N., Zelada-González, F., Banisch, T.U., Raible, F., Arendt, D., 2007. Vasa unveils Cnidaria. Development 138, 447–458. a common origin of germ cells and of somatic stem cells from the posterior growth Torras, R., Gonzalez-Crespo, S., 2005. Posterior expression of nanos orthologs during zone in the polychaete Platynereis dumerilii. Dev. Biol. 306, 599–611. embryonic and larval development of the anthozoan Nematostella vectensis. Int. Rebscher, N., Volk, C., Teo, R., Plickert, G., 2008. The germ plasm component vasa allows J. Dev. Biol. 49, 895–899. tracing of the interstitial stem cells in the cnidarian Hydractinia echinata. Dev. Dyn. Torras, R., Yanze, N., Schmid, V., González-Crespo, S., 2004. nanos expression at the em- 237, 1736–1745. bryonic posterior pole and the medusa phase in the hydrozoan Podocoryne carnea. Saffman, E.E., Lasko, P., 1999. Germline development in vertebrates and invertebrates. Evol. Dev. 6, 362–371. Cell. Mol. Life Sci. 55, 1141–1163. Tsunekawa, N., Naito, M., Sakai, Y., Nishida, T., Noce, T., 2000. Isolation of chicken vasa Sagawa, K., Yamagata, H., Shiga, Y., 2005. Exploring embryonic germ line development homolog gene and tracing the origin of primordial germ cells. Development 127, in the water flea, Daphnia magna, by zinc-finger-containing VASA as a marker. 2741–2750. Gene Expr. Patterns 5, 669–678. Voronina, E., Lopez, M., Juliano, C.E., Gustafson, E., Song, J.L., Extavour, C., George, S., Saito, T., Otani, S., Fujimoto, T., Suzuki, T., Nakatsuji, T., Arai, K., Yamaha, E., 2004. The Oliveri, P., McClay, D., Wessel, G., 2008. Vasa protein expression is restricted to germ line lineage in ukigori, Gymnogobius species (Teleostei: ) during em- the small micromeres of the sea urchin, but is inducible in other lineages early in bryonic development. Int. J. Dev. Biol. 48, 1079–1085. development. Dev. Biol. 314, 276–286. Salinas, L.S., Maldonado, E., Macías-Silva, M., Blackwell, T.K., Navarro, R.E., 2007. The Voronina, E., Seydoux, G., Sassone-Corsi, P., Nagamori, I., 2011. RNA granules in germ DEAD box RNA helicase VBH-1 is required for germ cell function in C. elegans. cells. Cold Spring Harb. Perspect. Biol. 3. doi:10.1101/cshperspect.a002774. Genesis 45, 533–546. Watanabe, H., Hoang, V.T., Mättner, R., Holstein, T.W., 2009. Immortality and the base Sato, K., Shibata, N., Orii, H., Amikura, R., Sakurai, T., Agata, K., Kobayashi, S., Watanabe, of multicellular life: lessons from cnidarian stem cells. Semin. Cell Dev. Biol. 20, K., 2006. Identification and origin of the germline stem cells as revealed by the 1114–1125. expression of nanos-related gene in planarians. Dev. Growth Differ. 48, 615–628. Weismann, A., 1883. Die Entstehung der Sexualzellen bei den Hydromedusen. Gustav Schröder, R., 2006. vasa mRNA accumulates at the posterior pole during blastoderm for- Fischer, Jena. mation in the flour beetle Tribolium castaneum. Dev. Genes Evol. 216, 277–283. Weismann, A., 1893. Germ-Plasm, a Theory of Heredity. Scribners, New York. English Seipel, K., Yanze, N., Schmid, V., 2004. The germ line and somatic stem cell gene Cniwi translation by W. N. Parker and H. Ronnfeldt. in the jellyfish Podocoryne carnea. Int. J. Dev. Biol. 48, 1–8. Wu, H.R., Chen, Y.T., Su, Y.H., Luo, Y.J., Holland, L.Z., Yu, J.K., 2011. Asymmetric localiza- Sekizaki, H., Takahashi, S., Tanegashima, K., Onuma, Y., Haramoto, Y., Asashima, M., tion of germline markers Vasa and Nanos during early development in the amphi- 2004. Tracing of Xenopus tropicalis germ plasm and presumptive primordial germ oxus Branchiostoma floridae. Dev. Biol. 353, 147–159. cells with the Xenopus tropicalis DAZ-like gene. Dev. Dyn. 229, 367–372. Yajima, M., Wessel, G.M., 2011. Small micromeres contribute to the germline in the sea Shirae-Kurabayashi, M., Nishikata, T., Takamura, K., Tanaka, K.J., Nakamoto, C., urchin. Development 138, 237–243. Nakamura, A., 2006. Dynamic redistribution of vasa homolog and exclusion of Ying, Q.L., Nichols, J., Chambers, I., Smith, A., 2003. BMP induction of Id proteins sup- somatic cell determinants during germ cell specification in Ciona intestinalis. presses differentiation and sustains embryonic stem cell self-renewal in collabora- Development 133, 2683–2693. tion with STAT3. Cell 115, 281–292.