Genetic dissection of the planarian reproductive system through characterization of Schmidtea mediterranea CPEB homologs.

Cytoplasmic polyadenylation is a mechanism of mRNA regulation prevalent in metazoan germ cells; it is largely dependent on Cytoplasmic Polyadenylation Element Binding proteins (CPEBs). Two CPEB homologs were identified in the planarian Schmidtea mediterranea. Smed-CPEB1 is expressed in ovaries and yolk glands of sexually mature planarians, and required for oocyte and yolk gland development. In contrast, Smed-CPEB2 is expressed in the testes and the central nervous system; its function is required for spermatogenesis as well as non-autonomously for development of ovaries and accessory reproductive organs. Transcriptome analysis of CPEB knockdown animals uncovered a comprehensive collection of molecular markers for reproductive structures in S. mediterranea, including ovaries, testes, yolk glands, and the copulatory apparatus. Analysis by RNA interference revealed contributions for a dozen of these genes during oogenesis, spermatogenesis, or capsule formation. We also present evidence suggesting that Smed-CPEB2 promotes translation of Neuropeptide Y-8, a prohormone required for planarian sexual maturation. These findings provide mechanistic insight into potentially conserved processes of germ cell development, as well as events involved in capsule deposition by flatworms.

1. Introduction

Sexual reproduction relies on assembly of functional gametes, as well as on processes that bring these together and support embryonic development. Genetic screens performed using invertebrate model organisms, such as Drosophila melanogaster and Caenorhabditis elegans, have revealed conserved genetic pathways that drive sexual reproduction. Molecular mechanisms for chromatin modification, mRNA regulation, and extrinsic signaling, have emerged from such studies as unifying themes in metazoan germline development (Kimble, 2011; Kimmins and Sassone-Corsi, 2005; Sasaki and Matsui, 2008; Seydoux and Braun, 2006; Voronina et al., 2011). Surveys performed in invertebrate model systems distinct from ecdysozoan phyla have been less explored, but begin to show promise in contributing to the body of scientific knowledge that drives advancements in reproductive medicine, as well as tempering of pests and parasites.

The phylum Platyhelminthes consists of free-living (planarians) and parasitic flatworms (trematodes, cestodes, and monogeneans). Planarians have long been studied for their ability to undergo whole-body regeneration and have emerged as model system in stem cell biology research (Elliott and Sanchez Alvarado, 2013; Newmark and Sanchez Alvarado, 2002; Rink, 2013; Shibata et al., 2010). These organisms can reproduce asexually by transverse fission, as well as sexually by hermaphroditic cross-fertilization (Newmark and Sanchez Alvarado, 2002; Newmark et al., 2008). Like mammals, planarians specify their germline through inductive mechanisms, whereas nematodes and flies utilize localized determinants (Extavour and Akam, 2003; Newmark et al., 2008; Wang et al., 2007). In fact, development of the planarian reproductive system occurs post-embryonically through differentiation of pluripotent stem cells, and can reoccur through a regenerative process upon injury (Chong et al., 2013; Collins et al., 2010; Newmark et al., 2008; Wang et al., 2007). An evolutionary novelty in the reproductive system of most flatworms is the presence of yolk glands (or vitellaria), which provide cellular material required for encapsulation and nutrition of embryos in ectolecithal eggs (Egger et al., 2015). This unique feature of flatworm reproduction makes yolk gland development a targetable mechanism for attenuating the dissemination and pathology of parasitic flatworms. Furthermore, sequence analyses have confirmed conservation of a fraction of human genes in planarians that are absent in genomes of ecdysozoan model systems (Sanchez Alvarado et al., 2002), and genetic contributions to development of the reproductive system can be studied by RNA interference (RNAi) (Wang et al., 2007). In sum, planarians are a valuable model for identifying genes involved in sexual reproduction of Platyhelminthes, as well as processes conserved across metazoan development.

Expression of Cytoplasmic Polyadenylation Element Binding protein (CPEB) homologs has been detected recently in the germline of the fluke Schistosoma mansoni and the non-parasitic flatworm Macrostomum lignano (Arbore et al., 2015; Cogswell et al., 2012; Lu et al., 2016). Members of the CPEB family of proteins are key regulators of mRNA during metazoan germline development. CPEB proteins can be divided in two subfamilies: CPEB1 (classic CPEBs) and CPEB2. CPEB1 orthologs regulate maternal mRNAs during oogenesis and early development (Castagnetti and Ephrussi, 2003; Christerson and McKearin, 1994; Hake and Richter, 1994; Lantz et al., 1994; Stebbins-Boaz et al., 1996; Tan et al., 2001). CPEB2 homologs are expressed in mouse testes (Kurihara et al., 2003) and required for spermatogenesis in nematodes (Luitjens et al., 2000) and flies (Xu et al., 2012). Representatives of both CPEB1 and CPEB2 subfamilies are expressed in the brain, where they regulate mRNAs involved in courtship, pain, and long-term memory formation (Chao et al., 2013; Huang et al., 2006; Keleman et al., 2007; Pai et al., 2013; Rouhana et al., 2005; Si et al., 2003; Wu et al., 1998; Zearfoss et al., 2008). The binding specificity, target repertoire, and mechanisms of regulation by CPEB1 orthologs have been analyzed in several model systems (Darnell and Richter, 2012; Fernandez-Miranda and Mendez, 2012; Groisman et al., 2001; Richter, 2007). Classical CPEBs mediate translational repression and activation by modulating changes in poly(A) tail length. Cytoplasmic Polyadenylation Elements (CPEs; UUUUA(U/A)) present in the 3′ untranslated regions (UTRs) recruit CPEB1 to target mRNAs (Barnard et al., 2004; Fox et al., 1989; Hake et al., 1998; Hake and Richter, 1994; Kim and Richter, 2006; Minshall et al., 1999, 2007; Novoa et al., 2010; Rouhana et al., 2005; Sheets et al., 1994). The binding specificity of CPEB2 subfamily members remains elusive. It has been reported that CPEB2 homologs bind to the same U-rich regulatory elements as CPEB1 (Afroz et al., 2014; Igea and Mendez, 2010; Novoa et al., 2010), but also to secondary RNA structures (Huang et al., 2006). Regulation of mRNA by CPEB2 has been shown to occur by poly(A) tail elongation (Igea and Mendez, 2010; Novoa et al., 2010; Ortiz-Zapater et al., 2012; Pique et al., 2008), poly(A) tail shortening (Hosoda et al., 2011), inhibition of translation independent of changes in poly(A) tail length (Huang et al., 2006), and nuclear functions (Kan et al., 2010).

In this study, we sought to uncover genetic mechanisms that regulate sexual reproduction of Platyhelminthes, utilizing the planarian Schmidtea mediterranea as a model system. We identified genes required for planarian oogenesis, spermatogenesis, and capsule formation, through the study of two S. mediterranea CPEB paralogs. We found that Smed-CPEB1 is expressed exclusively in female organs, and is required for oogenesis and yolk gland development. Smed-CPEB2 is expressed in testes and the central nervous system, and it is required for spermatogenesis as well as non-autonomously for development of ovaries and accessory reproductive organs. Transcriptome analyses of Smed-CPEB1 and Smed-CPEB2 knockdowns uncovered more than a hundred genes preferentially expressed in tissues of the reproductive system, a subset of which were functionally characterized by RNAi. Biochemical analyses support a model in which neuronal CPEB2 regulates expression of a signaling peptide required for sexual maturation. Collectively, these findings reveal molecular processes involved in metazoan germline development, as well as events specific to sexual reproduction of Platyhelminthes.

2. Results

2.1. Characterization of CPEB homologs in Schmidtea mediterranea

Two CPEB homologs (Smed-CPEB1 and Smed-CPEB2) were identified in the Schmidtea mediterranea Genome Database (Robb et al., 2008, 2015) by performing TBLASTN searches against human CPEB1-4 protein sequences (Fig. 1A,A′, and Supplementary Fig. 1A). Partial cDNAs were cloned based on genomic sequence and expression of mRNAs with corresponding sequence was verified by northern blot analysis (Supplementary Fig. 1B). Analysis of full-length ORF sequences obtained through 5′ and 3′ Rapid Amplification of cDNA Ends (RACE), revealed that both planarian CPEB homologs contain two RNA-recognition motifs (RRMs) and a Zinc Finger (ZnF) in their C-terminal ends, which is characteristic of CPEB protein architecture (Fig. 1B,C, and Supplementary Fig. 2A). Smed-CPEB1 (GenBank: KU990884; Steiner et al., 2016) shares highest sequence conservation with members of the classical CPEB subfamily of proteins with known roles in oogenesis (Fig. 1A). Whole-mount in situ hybridization analysis (WMISH) revealed Smed-CPEB1 expression exclusively in ovaries and yolk glands of sexually mature planarian hermaphrodites (Fig. 1D,D′). More specifically, ovarian Smed-CPEB1 expression was detected in oocytes (Steiner et al., 2016), which are identifiable by their large size and condensed chromosomes (Fig. 1E). Smed-CPEB2 (GenBank: KX074204) belongs to the CPEB2 protein subfamily (Fig. 1A). The C-terminal domain of this protein is 89% identical to the corresponding region of human CPEB2, which includes two RRMs and a ZnF (Fig. 1A′ and Supplementary Fig. 2B). CPEB2 expression was detected in the planarian brain and testes by WMISH (Fig. 1F,F′). Detailed expression analysis by fluorescent in situ hybridization (FISH) revealed that CPEB2 transcripts accumulate in spermatogonia, peak in spermatocytes, and vanish in the later stages of spermatogenesis (Fig. 1G).

Fig. 1

Expression of CPEB paralogs in sex-specific reproductive structures and central nervous system

(A) Neighbor-joining phylogenetic tree depicting the closer association of Smed-CPEB1 sequence with canonical CPEB subfamily proteins (magenta), and Smed-CPEB2 with members of the CPEB2/3/4 subfamily (blue). Phylogenetic analysis was performed using Clustal Omega under default parameters (Sievers et al., 2011). Abbreviations: Aplysia californica (Ac), Caenorhabditis elegans (Ce), Ciona intestinalis (Ci), Danio rerio (Dr), Drosophila melanogaster (Dm), Homo sapiens (Hs), Hydra vulgaris (Hv), Tribolium castaneum (Tc), Schmidtea mediterranea (Smed), and Xenopus tropicalis (Xt). HsMusashi is included as an outgroup. Scale bar represents 0.05 substitutions per amino acid position. (A′) Amino acid sequence identity shared between human and S. mediterranea CPEB RNA-binding domains. Highest identities for CPEB1 (pink) and CPEB2 (blue) are shown. (B-C) Human and planarian CPEB1 (B) and CPEB2 (C) protein architecture includes two RNA-recognition motifs (RRM; light gray) and a zinc finger (ZnF; dark gray). A nuclear export signal present in HsCPEB2 (NES; black) is partially conserved in Smed-CPEB2. HsCPEB1 contains a Proline Glutamic Acid Threonine (PEST) domain. (D-G) Wholemount (D, F) and fluorescent (E, G) in situ hybridization reveal expression of Smed-CPEB1 in ovaries (D) and yolk glands (D′) of S. mediterranea, whereas Smed-CPEB2 expression (F) was detected in the brain (arrows) and testes (F′). More specifically, Smed-CPEB1 mRNA was detected in oocytes (E) and Smed-CPEB2 in spermatogonia and spermatocytes. Scale bars =1 mm (D, F), 0.2 mm (D′, F′), and 20 μm (E, G). Nuclei are stained by DAPI (cyan) in (E) and (G).

Given the germline expression of planarian CPEBs, as well as known functions of characterized CPEBs in other organisms, we hypothesized that Smed-CPEB1 and Smed-CPEB2 would be required for oogenesis and spermatogenesis, respectively. To test this hypothesis, we subjected groups of six sexually mature planarians (~1.5 cm long and actively laying capsules) to weekly feedings of double-stranded RNA (dsRNA). A control group, subjected to dsRNA corresponding to bacterial ccdB sequence, produced egg capsules throughout the 8-week experiment (Supplementary Fig. 3). Planarians subjected to Smed-CPEB1 RNAi (CPEB1(RNAi)) or Smed-CPEB2 RNAi (CPEB2(RNAi)) produced capsules at rates comparable to the control group during the first two weeks of the RNAi treatment, produced less capsules than controls during the third and fourth weeks of RNAi treatment, and ceased producing capsules altogether during the last four weeks (Supplementary Fig. 3). At the end of the 8-week RNAi treatment, testes development was evaluated by labeling with the nuclear stain 4′, 6-Diamidino-2-Phenylindole, Dihydrochloride (DAPI), which facilitates visualization of organs with high cell density (e.g. copulatory apparatus, brain, pharynx), as well as sperm (Fig. 2A-D, A′-D′). The general anatomy of CPEB1(RNAi) appeared indistinguishable from that of control animals from this analysis (Fig. 2A,B). However, CPEB2(RNAi) displayed underdeveloped testes and copulatory apparatus (Fig. 2C). Similar, but more severe, defects were observed in nanos(RNAi) animals, which exhibit complete loss of germ cells (Fig. 2D; Wang et al., 2007, 2010). Closer analysis of testis anatomy by confocal microscopy of control and CPEB1(RNAi) revealed normal production of spermatozoa (Fig. 2A′,B′), which ultimately accumulated in the seminal vesicles (Fig. 2A,B). On the other hand, there were no spermatozoa or spermatids present in the under-developed testes of CPEB2(RNAi) (Fig. 2C′), suggesting that Smed-CPEB2 is required for meiotic entry or early progression. The defect in spermatogenesis observed in CPEB2(RNAi) is less severe than the complete loss of testes observed after nanos RNAi (Fig. 2D′; Wang et al., 2007). From these results we conclude that CPEB2 function is required for spermatogenesis and development of copulatory organs in S. mediterranea.

Fig. 2

Smed-CPEB1 is required during oogenesis and Smed-CPEB2 for general maturation of the reproductive system

(A-G) Analysis of reproductive anatomy of samples subjected to two-months of control (A, E), Smed-CPEB1 (B, F), Smed-CPEB2 (C, G) or nanos (D) RNAi. Abnormalities in development of testes, copulatory complex (penis papilla; white arrow), and accumulation of sperm in seminal vesicles, are apparent by DAPI staining in CPEB2(RNAi) (C) and nanos(RNAi) (D). Confocal microscopy revealed normal sperm production in control (A′) and CPEB1(RNAi) (B′), whereas CPEB2(RNAi) lacked spermatids and spermatozoa (C′). Testes were not detected in nanos(RNAi) (D′). (E-J) Detection of ovaries by pumilio in situ hybridization (ISH) (E-G; black arrows) and confocal analysis of DAPI-stained samples (H-J; yellow dashed lines) in controls (E, H) and CPEB1(RNAi) (F, I) samples. No ovaries were detected in CPEB2(RNAi) by either pumilio ISH (G) or confocal analysis of DAPI-stained samples (J). Oocytes (large cells with condensed chromosomes) present in ovaries of control samples (H) were not detected in ovaries of CPEB1(RNAi) (I). Detection of Smed-gH4 expression (green; H-J) was used as germline stem cell marker. Scale bars =1 mm (A-D), 50 μm (A′-D′ and H-J) and 0.5 mm (E-G).

The presence of ovaries in CPEB1(RNAi) and CPEB2(RNAi) planarians was assessed by WMISH analysis of a pumilio homolog preferentially expressed in ovaries of S. mediterranea (Fig. 2E; Zayas et al., 2005). Ovaries were readily detected in control and CPEB1(RNAi) samples (Fig. 2E,F). However, ovaries were not detected in CPEB2(RNAi) planarians (Fig. 2G). To verify these observations, ovary development was analyzed in detail by confocal microscopy of samples subjected to DAPI staining and FISH analysis of germinal histone H4 expression, which serves as a marker for oogonia (Wang et al., 2007; Rouhana et al., 2012). Indeed, ovaries were distinguishable in control and CPEB1(RNAi) planarians (Fig. 2H,I). Fully developed oocytes of approximately 20 μm in diameter and with condensed chromosomes were observed in control ovaries (Fig. 2H). However, CPEB1(RNAi) ovaries lacked oocytes and were filled with much smaller oogonia (Fig. 2I). Ovaries were not distinguishable in CPEB2(RNAi) samples, although presumptive oogonia were detected in the anatomical region of the ovaries (Fig. 2J). These results demonstrate that Smed-CPEB1 is required for oogenesis in S. mediterranea, whereas Smed-CPEB2 is required for development of the actual ovary. Altogether these results show that S. mediterranea CPEB homologs are required for germ cell development in gonads in which they are expressed. Additionally, Smed-CPEB1 is required for yolk gland development (Steiner et al., 2016), whereas Smed-CPEB2 is required for development of ovaries and accessory reproductive organs (e.g. copulatory apparatus) through non-autonomous mechanisms.

2.2. Characterization of the planarian reproductive system through analysis of CPEB1(RNAi) and CPEB2(RNAi) transcriptomes

The anatomical defects of CPEB1(RNAi) and CPEB2(RNAi) provided an opportunity to identify genes involved in sexual reproduction, particularly those expressed during later stages of oogenesis and spermatogenesis, as well as in accessory reproductive structures. Genes preferentially expressed in cell types absent in CPEB1(RNAi) (e.g. oocytes, yolk gland cells) and CPEB2(RNAi) (e.g. spermatids, spermatozoa, cells of the ovary and accessory reproductive organs) can be identified by comparing transcriptomes of control and CPEB knockdown planarians. Thus, we extracted total RNA from fertile control planarians and compared it with that of CPEB1(RNAi) and CPEB2(RNAi) animals of comparable size (Fig. 3A). Illumina reads from control (n=3 groups of 5 planarians; avg. reads/sample=20,605,364), CPEB1(RNAi) (avg. reads/sample=20,220,885), and CPEB2(RNAi) (avg. reads/sample=17,817,956) were mapped to a reference transcriptome composed of 55,949 contigs from S. mediterranea hermaphrodites (Rouhana et al., 2012). Comparison between control and CPEB1(RNAi) samples revealed 618 genes with statistically significant differences in expression ( > 2-fold; p < 0.05; excluding contigs with ≤ 2 avg. mapped reads). Of these, 333 were under-represented and 285 significantly over-represented in RNA from CPEB1(RNAi) (Fig. 3B, Supplementary Table 1). Genes with largest fold-change reduction in CPEB1(RNAi) included homologs of ras family proteins, tetraspan 1, tia1 and calponin, while over-represented genes were led by core 1 glycoprotein-n-acetylgalactosamine 3-beta, actin and ran (small nuclear GTP) -binding protein (Supplementary Table 1). Enrichment of nucleic acid binding, RNA binding (but not DNA binding), ion channel activity, and receptor activity GO category functions were observed among genes under-represented in CPEB1(RNAi) (Supplementary Fig. 4). GO categories for biological processes including transport, transmembrane transport, ion transport, calcium ion transport and intracellular signaling were enriched among CPEB1(RNAi) under-represented sequences (Supplementary Fig. 4). Over-represented transcripts in CPEB1(RNAi) were enriched in GO categories such as nucleotide binding, nucleic acid binding, DNA binding, RNA binding, GTP binding, and transcription factor activity; whereas enriched biological process GO categories included regulation of transcription, protein transport and response to stress (Supplementary Fig. 4). It is possible that transcripts with altered expression in CPEB1(RNAi) represent mRNA targets whose stability depends on deadenylation or polyadenylation events mediated by Smed-CPEB1. However, it is likely that under-represented transcripts correspond to genes expressed in cell types missing after CPEB1 knockdown. On the other hand, over-represented transcripts may represent a signaling response (or loss of negative feedback) due to the absence of the missing cell types after Smed-CPEB1 RNAi.

Fig. 3

RNA-seq analysis of CPEB1(RNAi) and CPEB2(RNAi) uncovers new markers of the planarian reproductive system

(A) Schematic experimental design of RNA-seq analyses highlighting differences between control (middle), CPEB1(RNAi) (left), and CPEB2(RNAi) reproductive anatomies. Yolk cells (yellow), ovaries (pink circles), ovaries without oocytes (open pink circles), oviducts (pink lines), testes (blue), and copulatory complex (orange glands; pink bursa). Dorsal (D) and ventral (V) views. (B-C) Plots displaying gene expression value means from control(RNAi) (x-axis) and CPEB1(RNAi) (y-axis) transcriptomes (B), as well as control(RNAi) (x-axis) vs. CPEB2(RNAi) (y-axis) (C). Genes with statistically significant (p value < 0.05) differences > 2-fold in transcript abundance are indicated (black dots). (D-I) WMISH showing markers of reproductive structures identified from RNAseq analysis of CPEB1(RNAi) (D, E) and CPEB2(RNAi) (F-I). Penis papilla (arrow in G) and oviducts (arrowheads in H) are indicated.

RNAseq analysis also revealed 1256 significantly under-represented and 808 over-represented transcripts from CPEB2(RNAi) samples (Fig. 3C, Supplementary Table 2). The genes with the largest increase in expression in CPEB2(RNAi) included Thrombospondin type 1 domain-containing protein, Aldolase, Porcupine-like protein and Cubitus interruptus homologs (Supplementary Table 2). Tetraspan 1, Surfactant b, and a conserved Plasmodium protein also identified in CPEB1(RNAi), were among the genes with the largest reduction in CPEB2(RNAi) (Supplementary Table 2). Enriched GO functional categories among under-represented sequences in CPEB2(RNAi) included hydrolase activity and peptidase activity, whereas the categories of DNA binding, nucleic acid binding and RNA binding were enriched among over-represented sequences (Supplementary Fig. 5). Biological processes enriched in CPEB2(RNAi) under-represented sequences include proteolysis and microtubule-based movement, whereas over-represented sequences included DNA integration, regulation of transcription, RNA-dependent DNA replication and G-protein coupled receptor signaling (Supplementary Fig. 5). Given the absence of ovaries, functional testes, and accessory reproductive organs in CPEB2(RNAi), we expect that many of the 1256 underrepresented sequences correspond to genes normally expressed in these structures. As with Smed-CPEB1, it is also possible that transcripts with altered expression in these samples represent Smed-CPEB2 target mRNAs, particularly in neurons, where Smed-CPEB2 expression was evident (Fig. 1F).

To establish markers for studying the anatomy of the planarian reproductive system, we performed WMISH analysis for transcripts with reduced abundance in CPEB1(RNAi) and/or CPEB2(RNAi). We first focused on fifty-two genes identified in Smed-CPEB1(RNAi) transcriptomes (Supplementary Table 3), which were prioritized based on largest fold-change of expression, detection of conserved domains through homology searches, and availability of cDNA clones prepared in our laboratory (Zayas et al., 2005). Expression in ovaries was verified by WMISH for 26 of these genes, which included a gelsolin homolog that labels ovarian cells facing the tuba (Fig. 3D). Four of the thirteen genes with the largest decrease in transcript abundance after CPEB1 RNAi displayed expression patterns reflective of yolk gland distribution (Supplementary Table 3). These included two previously characterized genes, surfactant b and synaptotagmin XV (Steiner et al., 2016), an uncharacterized protein conserved in the choanoflagellate Monosiga brevicollis (Fig. 3E), and a gene homologous to a protein identified in Plasmodium (Contig5529; PL030017B20H10; Supplementary Fig. 6). Reduction of all these yolk gland markers validates the previously reported requirement of Smed-CPEB1 function for yolk gland development (Steiner et al., 2016) and explains the loss of capsule production observed after CPEB1 RNAi (Supplementary Fig. 3).

Investigation of candidates identified from CPEB2(RNAi) transcriptome analysis included 132 sequences. Enriched expression in the testes was detected for 56 of these sequences (Fig. 3F, Supplementary Table 4). Markers were also identified for accessory reproductive organs such as the penis papilla (Cre-alp1 protein; Fig. 3G), the oviducts (a homolog of a hypothetical protein conserved in Schistosoma mansoni; PL04015A2A04; Fig. 3H), glands surrounding the gonopore (tetraspanin 66e; Fig. 3I), and yolk glands (see below). The high representation of testes markers (43.1%) is likely due to the abundance of this tissue in the planarian (Fig. 1F, 2A). Altogether, the transcriptomic analyses of CPEB1(RNAi) and CPEB2(RNAi) identified over a hundred genes specifically expressed in structures of the planarian reproductive system.

2.3. Identification of yolk gland genes required for capsule tanning

The capsules that house planarian embryos are produced independently from ovulation, fertilization or mating (Steiner et al., 2016). The capsule shell is initially formed in the genital atrium, which is located inside the gonopore posterior to the pharynx (Fig. 4A). These capsules are filled by hundreds of yolk cells that nurture the developing embryo (Martin-Duran et al., 2008; Shinn, 1993). Upon deposition, capsules undergo a “quinone tanning” process in which the proteinaceous shell hardens and darkens from yellow to dark-brown hues (Fig. 4B). Although materials required for capsule shell formation and tanning are believed to be contributed by yolk cells, little is known about the genetic factors contributing to this process.

Fig. 4

Genes required for capsule shell maturation and tanning

(A) S. mediterranea in the process of capsule deposition. (B) Time-course imaging of capsule tanning process, 0.5–16 h post-deposition. (C) Illustration of female reproductive anatomy with yolk gland distribution represented in yellow. (D-G) Detection of tanning factor-1 (D), surfactant b (E), tyrosinase (F), and c-type lectin (G) transcripts in yolk glands by WMISH. (H-L) Egg capsules produced by control (H), tanning factor-1(RNAi) (I), surfactant b(RNAi) (J), tyrosinase(RNAi) (K), and c-type lectin(RNAi) (L) > 1-month post-deposition. Scale bars=1 mm.

Genes expressed in patterns reflective of yolk gland distribution were identified in both CPEB1(RNAi) and CPEB2(RNAi) transcriptomes (Fig. 4C-G; Supplementary Tables 3 and 4). From these, four were found to be required for capsule shell pigmentation (Fig. 4H-L). One encodes an uncharacterized protein, which we named Smed-tanning factor-1 (tan-1) (Fig. 4D). The other three genes encode homologs of Surfactant B, Tyrosinase, and C-type lectin (Fig. 4E-G). Expression of tyrosinase was also strongly detected posterior to the copulatory complex (Fig. 4F). Capsules produced by tan-1(RNAi) animals developed a dark reddish hue comparable to that of normal capsules 4–8 h post-deposition (Fig. 4B, I), but failed to reach the dark brown color observed in normal capsules (Fig. 4H). Capsules from surfactant b(RNAi) and tyrosinase(RNAi) developed more severe coloration defects, maintaining yellow to orange hues comparable to those of normal capsules 30 min to 2 h post-deposition (Fig. 4J, K). c-type lectin(RNAi) capsules displayed the most severe shell coloration defect, manifested with an overall milky color and a few dark speckles (Fig. 4L). The capsule coloration defects observed for all four genes lasted for at least three months. Additionally, no hatchlings emerged from any of the capsules that displayed coloration defects during this tme (tan-1, n=12; surfactant b, n=18; tyrosinase, n=8; and c-type lectin, n=12), suggesting that “tanning” and/or processes that lead to capsule maturation are essential for planarian embryonic development. These processes may also play a role in reproduction of parasitic flatworms, since a subset of these genes is conserved in schistosomes and/or tapeworms (Supplementary Table 5).

2.4. A Synaptotagmin homolog required for capsule shell assembly

Another gene with reduced transcript levels in both CPEB1(RNAi) and CPEB2(RNAi) encodes a synaptotagmin-like family member (Smed-synaptotagmin XV or sytXV). Detection of sytXV expression by WISH (Fig. 5A) is confined to the ovaries (Fig. 5A′) and yolk glands (Fig. 5A″). More specifically, ovarian sytXV expression has been shown in oocytes (Steiner et al., 2016). Using Smed-CPEB1 as a marker for oogenesis and yolk gland development, analyses by WMISH showed that neither process is affected in sytXV(RNAi) (Fig. 5B,C). However, analysis of capsule production by sytXV(RNAi) animals revealed that they produce defective structures during deposition (Fig. 5D,E). The aberrant capsules produced by sytXV(RNAi) contained a thin membrane holding the yolk cells together, but this fragile structure eventually collapsed without yielding hatchlings. These results suggest that sytXV function has a direct role in capsule assembly, and possibly mediates exocytic release of yolk cell material required for shell formation.

Fig. 5

Identification of genes required for oogenesis and capsule formation

(A) WMISH detects sytXV expression in ovaries (A′) and yolk glands (A″) of S. mediterranea. (B-C) Assessment of normal oocyte and yolk gland development using Smed-CPEB1 as a marker in control(RNAi) (B) and sytXV(RNAi) (C). (D-E) Images of normal capsules produced by control(RNAi) (D) and shell-less capsules produced by sytXV(RNAi) (E) captured by dark field microscopy. (F) eIF4e-like mRNA detection in testes (F′), ovaries (arrows; F″), and presumably gut goblet cells (F″ and F″). (G-H) Confocal sections of control (G) and eIF4e-like(RNAi) (H) reveal a decrease in oocyte number (large cells with condensed chromosomes visualized by DAPI (cyan)) and an increase in cells expressing gelsolin (magenta) in ovaries (dashed line) of eIF4e-like(RNAi) (1.5 vs. 4.7 oocytes/ovary, and 77.0 vs. 25.7 gelsolin (+) cells/ovary, of eIF4e-like(RNAi) vs. controls, respectively; p < 0.05 two-tailed unpaired Student′s t-test, n ≥ 3 biological replicates per group). Scale bars=1 mm (A-F) and 100 μm (G-H).

2.5. A germline eIF4E paralog is required for oogenesis in S. mediterranea

Functional analysis of a eukaryotic translation initiation factor 4E homolog (Smed-eIF4E-like) underrepresented in CPEB1(RNAi) transcriptomes revealed a pleiotropic phenotype. Smed-eIF4E-like expression was detected in ovaries, testes, and a subset of gut cells, presumably goblet cells (Fig. 5F-F″). Analysis of comparably sized samples revealed that eIF4E-like(RNAi) ovaries contained reduced oocyte numbers compared to controls (1.5 oocytes/ovary vs. 4.7 oocytes/ovary, p < 0.05 two-tailed unpaired Student′s t-test) (Fig. 5G,H). On the other hand, the number of non-oogenic cells in the ovary, visualized by gelsolin FISH, was higher in eIF4E-like(RNAi) than in control ovaries (Fig. 5G,H; 77.0 gelsolin (+) cells/ovary vs. 25.7 gelsolin (+) cells/ovary, p < 0.05 two-tailed unpaired Student′s t-test). Spermatogenesis defects were not observed in eIF4E-like(RNAi) animals compared with control samples of the same size (data not shown). However, a significant reduction in capsule production (Supplementary Fig. 7), along with a growth defect (avg. size post-fixation=0.81 cm vs. 0.51 cm, p < 0.01 two-tailed unpaired Student′s t-test) were observed in eIF4E(RNAi) after prolonged (≥ 4 weeks) RNAi treatments. We conclude that the function of this eIF4E paralog is required for oocyte development and hypothesize that the reduction in capsule production may result from growth defects observed after prolonged Smed-eIF4E-like RNAi.

2.6. Identification of conserved factors required in late stages of planarian spermatogenesis

Given the defects in spermatogenesis observed in CPEB2(RNAi) testes (Fig. 2C), we hypothesized that genes under-represented in Smed-CPEB2(RNAi) transcriptomes include factors required for sperm elongation and function. We tested 57 candidate genes with validated testis expression (Supplementary Table 4) for spermatogenesis defects after RNAi. Defects in maintenance of sperm production by sexually mature planarians were visualized by DAPI staining and confocal microscopy at the end of a 6-week RNAi treatment. Six genes were found to be required for proper sperm development (Fig. 6A-F). From these genes, Smed-rap55 (RNA-associated protein 55 (RAP55/LSM14)) was previously reported to function in planarian spermatid elongation (Wang et al., 2010), thus validating our approach. The five other genes found to be required for spermatid elongation are planarian homologs of cysteine/histidine rich-1 (Smed-cyhr-1), the SWI/SNF chromatin remodeling complex component SMARCB1, polyubiquitin2, as well as t-complex protein alpha (Smed-cct-1) and t-complex protein delta (Smed-cct-4) (Fig. 6B-F). Together with the identification of factors required for capsule development and oogenesis, this study provides a set of candidate genes with functions specific to sexual reproduction of flatworms (yolk gland genes), as well as genes with conserved roles in metazoan germline development.

Fig. 6

Genes required for spermatid elongation

Confocal images of testes stained with DAPI from control (A) display normal sperm production and testis morphology, whereas cyhr-1b (B), SMARCB1 (C), polyubiquitin2 (D), cct-1 (E), and cct-4 (F) RNAi lead to spermatid elongation defects. Scale bars=50 μm.

2.7. CPEB2(RNAi) leads to decreased Neuropeptide Y-8 levels

The fact that Smed-CPEB2 RNAi affected the development of tissues in which its expression was not detected (e.g., ovaries, copulatory apparatus and yolk glands), led us to hypothesize that regulation of neuronal mRNAs by CPEB2 is required for development of these structures. A natural target for CPEB2 under this model is neuropeptide Y-8 (npy-8), which is required for sexual maturation; knockdown of this gene closely phenocopies CPEB2(RNAi) (Collins et al., 2010). Analysis of the npy-8 mRNA 3′-UTR revealed the presence of sequence sufficient for human CPEB2 binding in vitro (UUUUA; Afroz et al., 2014) (Fig. 7A), suggesting the possibility for direct Smed-CPEB2 binding. Furthermore, co-expression analysis by FISH demonstrated that neurons expressing npy-8 also contain Smed-CPEB2 transcripts (Fig. 7B-B″ ). We tested whether npy-8 expression levels were affected by CPEB2 RNAi by northern blot as well as RT-qPCR, and found npy-8 mRNA was ~45% less abundant in total RNA extracts of CPEB2(RNAi) compared to control(RNAi) or CPEB1(RNAi) samples (Fig. 7C,D). Analysis by western blot using anti-NPY-8 antibodies (Saberi et al., in revision) revealed that the NPY-8 peptide present in extracts of control and CPEB1(RNAi) planarians was dramatically reduced in CPEB2(RNAi) extracts (Fig. 7E). Thus, inhibition of Smed-CPEB2 expression leads to a severe reduction in NPY-8 peptide levels. These results support the model that Smed-CPEB2 regulates sexual maturation non-autonomously through stimulation of npy-8 mRNA translation in neurons.

Fig. 7

Neuropeptide Y-8 abundance decreases upon Smed-CPEB2 RNAi

(A) 3′UTR sequence of npy-8 mRNA (GenBank: BK007010) shows a CPE (gray oval) and potential cleavage and polyadenylation hexanucleotide (white oval). ORF is presented (not to scale; red box) with stop codon. (B) Detection of Smed-CPEB2 mRNA (magenta) in neurons expressing npy-8 (green; B′). DAPI-stained nuclei (gray) and position of neuron expressing npy-8 (arrowhead) are shown. (B″) Merged image. Scale bar=20 μm. (C) Northern blot analysis of npy-8 mRNA in total RNA extracts of control, CPEB1(RNAi), and CPEB2(RNAi). DNA size markers shown in left. (D) RT-qPCR analysis of npy-8 mRNA in total RNA from biological triplicates of control, CPEB1(RNAi), and CPEB2(RNAi) normalized to β-tubulin mRNA. (E) Western blot analysis of NPY-8 neuropeptide in protein extracts of control, CPEB1(RNAi), and CPEB2(RNAi). Position of size markers is shown. Non-specific 15 kD signal (asterisk) serves as endogenous loading control.

3. Discussion

To examine the contributions of cytoplasmic polyadenylation to germ cell biology, we characterized the expression and function of CPEB homologs in S. mediterranea. We showed that Smed-CPEB1 is expressed in the ovaries and yolk glands of sexually mature planarians, while Smed-CPEB2 is expressed in the testes and central nervous system. Smed-CPEB1 is required for planarian oogenesis and yolk gland development, whereas Smed-CPEB2 is required for spermatogenesis as well as for sexual maturation through a non-autonomous mechanism involving regulation of neuropeptide y-8 expression. This study complements previous work (Chong et al., 2011; Wang et al., 2010) in providing a collection of markers for the study of the planarian reproductive system. We identified markers for different components of the planarian reproductive system, including but not limited to the testes, distinct ovarian cell types, the oviducts, the penis papilla, and yolk glands.

Functional analysis of candidates with validated expression in ovaries, testes, and yolk glands, led to the identification of 11 genes required for germline development and/or sexual reproduction in S. mediterranea. These include genes encoding proteins with uncharacterized domains (i.e. tanning factor-1), and genes with characterized homologs not previously known to play a role in germline development (e.g. SMARCB1 and cyhr1). We identified two planarian genes required for spermatid elongation (cct-1 and cct-4) that encode components of the chaperonin containing TCP-1 (CCT) complex. The CCT complex is highly conserved in eukaryotes and key for actin and tubulin folding (Valpuesta et al., 2002). SMARCB1 belongs to the family of switch/sucrose nonfermentable (SWI/SNF) chromatin remodeling complex proteins, and is the second component of this complex shown to be required for planarian spermatogenesis. Wang et al. (Wang et al., 2010) previously showed that the central catalytic ATPase of the SWI/SNF complex, the brahma-related gene 1 homolog (Smed-Brg1; or SMARCA4) is required for spermatid elongation. Mammalian orthologs of these genes have broad expression patterns, but their analysis is S. mediterranea revealed functional requirements during spermatogenesis, which demonstrates one advantage of using different models systems for studying germline development. Recent evidence shows that the contributions of SWI/SNF factors uncovered in planarian spermatogenesis studies hold true in mammals, as knockout of Brg1 in mouse testes leads to arrest of sperm development at midpachytene stage (Kim et al., 2012; Wang et al., 2012). Similarly, current tools for tissue-specific knockout studies in mice can facilitate analysis of mammalian orthologs of other genes characterized in planarians, and determine the conservation of molecular pathways contributing to metazoan germline development.

Twenty-six genes with ovarian expression were identified through RNAseq analysis of CPEB1(RNAi). Amongst these, an eIF4E homolog was found to be required for oogenesis (Fig. 5H). This result corroborates reports from the study of a homolog in nematodes (ife-1; Henderson et al., 2009). Interestingly, an ovary-specific isoform of eIF4E (eIF4Eb) was identified as a component of CPEB-containing complexes in Xenopus oocytes (Minshall et al., 2007). Contrary to canonical eIF4E, which binds the 5′-cap and facilitates recruitment of ribosomes, eIF4Eb appears to inhibit translation of mRNAs (Kubacka et al., 2015; Minshall et al., 2007). Indeed, our experiments show that both CPEB1 and eiF4e-like are required for development of oocytes in S. mediterranea, so these may function as a complex in this context.

Like Smed-CPEB1, expression of sytXV is restricted to yolk glands and ovaries (Fig. 5A), more specifically oocytes (Steiner et al., 2016). However, sytXV RNAi did not lead to defects in oogenesis or yolk gland development (Fig. 5C), instead it resulted in a severe capsule shell formation defect (Fig. 5E). This defect is attributed to disruption of sytXV expression in yolk cells, since it has been shown that oocytes are dispensable for normal capsule formation (Steiner et al., 2016). Synaptotagmin proteins play a role in regulating exocytosis that has been primarily studied in neurons (Sudhof, 2002). This family of proteins is characterized by the presence of an N-terminal transmembrane region and two C2 domains. Smed-SytXV is more similar to Sytl4 and Syt15 than any other mouse protein, it contains two C2 domains but does not have a transmembrane region recognizable by domain prediction software (i.e. TMpred, TMHMM). Thus the particular phenotype of Smed-sytXV(RNAi) could be explained by a reduction of SytXV-mediated secretion of yolk factors required for shell matrix formation. Indeed, yolk cells contain granules with materials that are released by exocytosis and form the capsule shell (or “wall”; Shinn, 1993), and the identification of sytXV function in this process paves the way for dissecting the molecular events involved in capsule shell formation.

Four additional genes required for development of the capsule shell were identified (Fig. 4). Yolk cells inside the developing capsule are thought to secrete proteins rich in 3,4-dihydroxyphenyl-L-alanine (DOPA) and enzymes that oxidize DOPA into o-quinone, which ultimately polymerizes with other proteins to harden and pigment the shell (Gremigni and Domenici, 1974; Ishida and Teshirogi, 1986; Marinelli, 1972; Nurse, 1950). Indeed, all four genes found to contribute to the tanning process are expressed in yolk glands, providing additional evidence for the contribution of yolk cells to shell development. Among these, Tyrosinase is known to catalyze polymerization of proteins in vitro and to be involved in the quinone tanning process through DOPA modifications of tyrosine-rich proteins in Platyhelminthes and insects (Ishida and Teshirogi, 1986; Waite, 1983). It is not clear how other yolk cell gene products influence the quinone tanning process, but it seems clear that the molecular events associated with this process are required for planarian embryonic development, since capsules with pigmentation defects failed to yield any hatchlings. Identification of proteins required for capsule shell formation may prove valuable in medical efforts to attenuate egg (capsule) production by parasitic flatworms.

CPEB homologs are key factors for post-transcriptional regulation during germline development and in neurons. The RNA-binding domain of human CPEB2 subfamily members shares ≥ 85% amino acid sequence identity with Smed-CPEB2 (Fig. 1A′, Supplementary Fig. 2), suggesting that mechanisms for target recognition and regulation are likely conserved between planarians and mammals. Although modulation of neuropeptide expression is not a known function of mammalian CPEB homologs, regulation of a conserved handful of brain mRNAs by cytoplasmic polyadenylation has been reported in flies, Aplysia, Xenopus, and mice (Gerstner et al., 2012; Ivshina et al., 2014; Rouhana et al., 2005). Future research will uncover the identity and fate of mRNAs regulated by Smed-CPEB2 in neurons and testes, which will reveal additional pathways involved in sexual maturation cell autonomously (regulation of proliferation and differentiation in sperm precursors) and non-autonomously (by modulation of neuronal signaling).

4. Materials and methods

4.1. Planarian culture

Laboratory lines of hermaphroditic Schmidtea mediterranea (Zayas et al., 2005) were used for all experiments and maintained in 0.75X Montjuïc salts as per (Wang et al., 2007). Animals were fed calf liver and starved for a week prior to experimentation. Capsules were collected weekly, maintained and monitored in 0.75X Montjuïc salts at 18 °C for ≥ 2 months post-deposition.

4.2. Whole-mount in situ hybridization and DAPI staining

Sample fixation for WMISH and DAPI staining was performed as per (King and Newmark, 2013) with modifications for large planarians that include 7–10 min treatment with 10% N-acetyl cysteine in PBS, and a 1.5 h fixation at 4 °C using 4% formaldehyde in PBSTx (PBS; 0.3% Triton-X). Hybridization of riboprobes was extended to ≥ 36 h. Colorimetric signal development was performed as per (Pearson et al., 2009). Nuclei were stained by incubation in PBSTx solution containing 5 μg/mL 4′,6-diamidino-2-phenylindole (DAPI) overnight at 4 °C, followed by four washes in PBSTx. Low-magnification images were captured with Zeiss V.12 and V.16 SteREO units, and high-magnification images by confocal microscopy.

4.3. RNA-interference

RNAi was performed as per (Rouhana et al., 2013). Briefly, groups of six to seven planarians were fed to satiation with liver solution containing dsRNA (100 ng/μL concentration) every 5–7 days. DsRNA of bacterial ccdB gene sequence was used as a negative control.

4.4. RNAseq analysis

Total RNA from biological triplicates of control(RNAi), CPEB1(RNAi) and CPEB2(RNAi) planarians ( > 1.5 cm) was extracted using TRIzol (invitrogen), treated with RQ1 DNase (Promega) for 10 min at room temperature, extracted again with TRIzol, dissolved in RNase-free water, and submitted to the Roy J. Carver Biotechnology Center (Univ. of Illinois at Urbana-Champaign) for library preparation (Illumina TruSeq RNA Sample prep kit) and sequencing (Illumina HiSeq. 2000 version 5 chemistry and analysis pipeline 1.8). Sequence reads were mapped to a S. mediterranea hermaphrodite reference transcriptome (Rouhana et al., 2012; “uc_Smed_v2” in PlanMine (Brandl et al., 2016)) and quantified using CLC genomics workbench (Qiagen) under default settings as in (Rouhana et al., 2012). The sequence dataset generated from RNAseq is publicly available (BioProject ID PRJNA319057).

4.5. Northern blot and RT-qPCR

Riboprobes corresponding to partial Smed-CPEB1 and Smed-CPEB2 ORFs, including RRMs and ZnF domains, were synthesized using T3 polymerase from pJC53.2/Smed-CPEB1 and pJC53.2/Smed-CPEB2, and used in northern blots as per (Miller and Newmark, 2012). Methods used for quantification of npy-8 mRNA by northern blot and RT-qPCR have been described (Collins et al., 2010).

4.6. 5′ and 3′ Rapid Amplification of cDNA Ends (RACE)

Full-length cDNAs were obtained using a SMARTer® RACE 5′/3′ Kit (Clontech Laboratories) as per manufacturer protocol. The primers used were: Smed-CPEB1 5′RACE, 5′-GATTACGCCAAGCTTCGTGTTGAGAATT AAATGTTTGTTG-3′; Smed-CPEB1 3′RACE, 5′-GATTACGCCAAG CTTAACTCCAGCCTTAAAGACATTTG-3′; Smed-CPEB2 5′RACE, 5′-GATTACGCCAAGCTTATTGGTGGAATTTATCGAATCATAG-3′; and Smed-CPEB2 3′RACE, 5′- GATTACGCCAAGCTTAGCTATTGAATTGGCAATGATTATG-3′.

4.7. NPY-8 Extraction and Western Blot

Planarian neuropeptides were extracted using a modification of published methods (Sturm et al., 2010) prior to detection by western blot. Endogenous proteases were inactivated by incubating worms at 80 °C for 1 min and samples were frozen on dry ice, transferred to a chilled glassteflon or glass-glass (small clearance) homogenizer and homogenized by 20 strokes in 1 mL acidified methanol (methanol, acetic acid, and water at 90:9:1;) for every 5 worms. The homogenate was stirred at 4 °C for an hour and centrifuged at > 14,000g for 20 min. The supernatant was frozen in liquid nitrogen, vacuum-dried overnight, and the pellet was solubilized in 100 μL of 1X lithium dodecyl sulfate (LDS) sample buffer containing 100 mM DTT by incubation at 70–80 °C for 10 min. 10 μL of the supernatant was loaded on a NuPAGE Novex 12% Bis-Tris gel (Invitrogen), resolved in 2-ethanesulfonic (MES) running buffer, and transferred onto a presoaked Immobilon-PSQ membrane in 2X NuPAGE Transfer Buffer (Invitrogen) containing 20% methanol at 10 V for 40 min. After transfer, the membrane was air-dried for 2 h and incubated in 5% BSA, 1% casein, and affinity-purified anti-NPY-8 antibody (1:2000 dilution) in PBS containing 0.05% Tween-20 (PBSTw) for 4 h at 4 °C. After three PBSTw washes, incubation with HRP-conjugated anti-rabbit antibodies (Jackson Laboratories; 1:10,000), and additional washes in PBSTw, the membrane was treated with ECL Western Blot Detection Reagent (Amersham) and imaged with a FluorChem Q system (Alpha Innotech).