1 Introduction

Animal germline cells are specified either through zygotic induction or cytoplasmic inheritance (see Extavour and Akam 2003; Extavour 2007; Quan and Lynch 2016 for a review). Zygotic induction takes place in mid- or late embryogenesis; it requires cell-to-cell signaling and leads to the acquisition of germline fate de novo. In contrast, cytoplasmic inheritance (called also maternal provision) involves formation of a specific, asymmetrically localized oocyte region, termed the germ plasm. This region contains maternally provided germline determinants (mRNAs, proteins) that are capable of inducing germline fate in a subset of embryonic cells. In insects, this region of the oocyte cytoplasm (ooplasm) is called the pole plasm or oosome. Interpretation of published data in a phylogenetic context leads to the suggestion that zygotic induction is more common among animal taxa and most probably represents the ancestral mode (mechanism) of germline specification (ancestral condition), whereas cytoplasmic inheritance evolved secondarily (derived condition) and independently in several animal lineages (Extavour and Akam 2003; Quan and Lynch 2016).

The best understood example of cytoplasmic inheritance comes from the studies of the fruit fly, Drosophila melanogaster. In this model species, the maternally derived germline determinants become localized to the posterior pole where they participate in the formation of the pole plasm (see Sect. 5.3). During early embryogenesis, the embryonic (blastoderm) nuclei that interact with the pole plasm (germline determinants) become cellularized before the other blastoderm nuclei. These cells, called pole cells, give rise to primordial germ cells (PGCs). Classical experimental studies clearly indicate that the pole plasm and the pole cells are both necessary and sufficient for PGC specification (see Mahowald 2001 for a review). Recent data suggest that the mode of germline specification described in the fruit fly is not ancestral among insects. So far, the asymmetrically located pole plasm and pole cells have been described only in taxa nested within Holometabola (insects undergoing complete metamorphosis) and three species (the book louse, Liposcelis divergens; mullein thrips, Haplothrips verbasci; and pea aphid, Acyrthosiphon pisum; see Goss 1952; Heming 1979; Lynch et al. 2011; Ewen-Campen et al. 2013; Lin et al. 2014 for further details), belonging to Paraneoptera, the sister group of Holometabola (Kristensen 1981). These observations led to the idea that the cytoplasmic inheritance is a derived feature of Holometabola (Lynch et al. 2011) or Holometabola and its sister taxon, Paraneoptera (Ewen-Campen et al. 2013) (Fig. 5.1, triangle and diamond, respectively). Although this idea seems to be well established, the evolution of germline specification among insects must have been much more complicated, because in several holometabolous lineages, cytoplasmic inheritance has been secondarily lost (Fig. 5.1, taxon names marked with asterisk) and replaced (secondarily) by zygotic induction (discussed in Lynch et al. 2011; Quan and Lynch 2016).

Fig. 5.1
figure 1

Simplified cladogram of Paraneoptera + Holometabola (=Phalloneoptera) (based on hypotheses of Henning (1981) and Kristensen (1995). The distribution of ovary types is superimposed: black box indicates putative origin of meroistic ovary; red thick lines indicate reversion to neopanoism. Evolutionary origin of the cytoplasmic inheritance as suggested by Ewen-Campen et al. (2013) is indicated by a green diamond, while that proposed by Lynch et al. (2011) and Quan and Lynch (2016) by a blue triangle. Strategy of germline specification operating in a given taxon (order, family, or species) is indicated by a color of its name and marked with asterisk, black dot, or #. Taxon names in blue (not marked), cytoplasmic inheritance; names in red (marked with asterisk), zygotic induction; while those in black (marked with black dot), both strategies: in some species cytoplasmic inheritance in other zygotic induction. Names in green (marked with #)—cytoplasmic inheritance suggested or independently derived; boxes with ?—data not conclusive. Ph Phalloneoptera, P Paraneoptera, H Holometabola, N Neuropterida, M Megaloptera, Of Osmylus fulvicephalus, Sm Sialis mitsuhashii, Tc Tribolium castaneum, Ao Acanthoscelides obtectus, Cm Callosobruchus maculatus, Dm Drosophila melanogaster, Cp Culex pipiens, Ag Anopheles gambiae, Pc Panorpa communis, Bm Bombyx mori, Pg Pectinophora gossypiella, Am Apis mellifera, Nv Nasonia vitripennis, Co Cosmoconus meridionator, Rp Rhodnius prolixus, Ap Acyrthosiphon pisum (literature cited in the text)

Full size image

Intriguingly, both classical histological and recent data additionally indicate that cytoplasmic inheritance is, as a rule, confined to those insect taxa that are characterized by meroistic ovaries. As evolution (anagenesis) of insect ovaries is rather complicated and often not fully understood, we will briefly review present knowledge of insect ovary types in the next section.

2 Morphology of Insect Ovarioles from a Phylogenetic Perspective

Insect ovaries consist of several elongated tube-shaped elements, termed ovarioles (see Büning 1994 for a review). The ovarioles are usually composed of three well-defined regions: the terminal filament, the germarium, and the vitellarium. The terminal filament is a stack of flat, disk-shaped somatic cells, oriented perpendicular to the long axis of ovariole. The germarium contains dividing and/or differentiating oogonial cells, whereas the vitellarium consists of linearly arranged, sequentially growing ovarian follicles (reviewed in Büning 1994). Two basic morpho-functional categories of insect ovarioles are traditionally recognized, panoistic and meroistic (Brandt 1874).

In the vast majority of insects, oogenesis starts with the formation of germline cysts composed of several interconnected sibling cells (reviewed in Büning 1994). The processes underlying cyst formation have been extensively studied in a model species, Drosophila melanogaster and in dozens of closer and more distant relatives of the fruit fly, including flies, beetles, wasps, butterflies, lacewings, hemipterans (true bugs, aphids, coccoids, etc.), earwigs, and lice (see Büning 1994; Klag and Bilinski 1994; Kubrakiewicz 1997; Bilinski 1998; Pyka-Fosciak and Szklarzewicz 2008; Tworzydlo et al. 2010 for further details). These studies have shown that initial stages of cyst morphogenesis are similar (evolutionary stable) in all investigated species. In contrast, the fate of interconnected cyst cells in subsequent stages of oogenesis is remarkably different in panoistic versus meroistic ovarioles. In the panoistic ovarioles, the germline cysts split into functionally equivalent cells that become definitive oocytes (Pritsch and Büning 1989; Gottanka and Büning 1990). In the meroistic ovarioles, the cyst cells remain joined; one of them differentiates into the oocyte, and the remaining cells form the nurse cells. The main function of nurse cells is the synthesis and subsequent transport of the macromolecules and organelles to the oocyte cytoplasm, while the oocyte nucleus, as a rule, remains transcriptionally quiescent. Depending on spatial organization of the ovariole as well as relationships between the oocyte(s) and the nurse cells, two distinct categories of meroistic ovarioles are usually recognized: polytrophic and telotrophic. In the polytrophic ovariole, each oocyte is accompanied by its own group of nurse cells, whereas in the telotrophic ovariole, all nurse cells are retained in the germarium, thereby forming a trophic chamber.

Analysis of the distribution of the ovary/ovariole categories in various insect lineages in the light of phylogenetic hypotheses of Henning (1981) and Kristensen (1995) leads to the following conclusions:

  • Panostic ovaries represent a plesiomorphic character (ancestral condition) inherited from the common hexapod ancestor. This ovarian category is characteristic for all basally branching taxa.

  • Meroistic-polytrophic ovaries evolved, from the panoistic ones, in the common ancestor of Dermaptera and Phalloneoptera (Paraneoptera + Holometabola) (Fig. 5.1, black box).

  • In some groups (Thysanoptera, Siphonaptera, Megaloptera: Corydalidae), meroistic ovaries have been reversed to the panoistic state (Fig. 5.1, red thick lines). These ovaries are characterized by the (secondary) loss of nurse cells; they are termed neopanoistic to indicate evolutionary distinctness from primary panoistic ones (see Stys and Bilinski 1990; Büning 1993, 1994 for further details).

Figure 5.1 shows a simplified cladogram of Paraneoptera + Holometabola (=Phalloneoptera). The distribution of ovariole types as well as the phylogenetic pattern of a single acquisition and multiple losses of the pole plasm and pole cells are superimposed on the cladogram. Two scenarios of the evolutionary origin of the cytoplasmic inheritance as suggested by Lynch et al. (2011) and Ewen-Campen et al. (2013) are indicated on the cladogram by triangle and diamond, respectively. Analysis of the superimposed events suggests that at least some of the multiple losses of the cytoplasmic inheritance (germ plasm) are associated with the loss of nurse cells, i.e., a reduction to the neopanoistic condition. Due to the sparse sampling, the proposed scenario/s are far from being complete and understood; therefore, we present new and unpublished data regarding ooplasm differentiation in the representatives of two non-holometabolous orders, Thysanoptera and Dermaptera (see Sects. 5.5 and 5.6).

3 The Origin and Assembly of Pole Plasm in Drosophila

It is well established that in Drosophila, pole plasm constituents have maternal origin and are synthesized during oogenesis, well ahead of embryogenesis (reviewed in Mahowald 2001; Becalska and Gavis 2009; Rangan et al. 2009; Lehmann 2016). The origin and interactions of the pole plasm components are both spatially and temporally regulated during oogenesis; therefore, a brief description of the fruit fly ovary architecture and functioning is necessary.

The Drosophila ovaries are composed of 12–14 canonical meroistic-polytrophic ovarioles (King 1970, see also Sect. 5.2) that comprise multicellular complexes called ovarian follicles or egg chambers. Each follicle is composed of a centrally located cyst of germline cells surrounded peripherally by somatic cells, the follicular cells. The germline cysts originate from four consecutive mitotic divisions of a cystoblast, a specialized stem cell-derived progenitor cell. Because these divisions are not followed by complete cytokineses, the resulting 16 cells are interconnected by intercellular bridges, also known as ring canals. Within each 16-cell cyst, one cell adopts an oocyte fate, while the remaining 15 differentiate into nurse cells (for a detailed description of the germline cyst formation and development, see de Cuevas et al. 1997; Bastock and St Johnston 2008). During egg chamber development, nurse cells become polyploid and highly transcriptionally active. In contrast, the oocyte is largely transcriptionally quiescent, enters meiosis, and ultimately becomes a haploid egg cell. During late oogenesis, the nurse cells undergo apoptosis and rapidly pass most of their cytoplasmic contents into the oocyte in a process called nurse cell dumping (Mahajan-Miklos and Cooley 1994; Cavaliere et al. 1998). Subsequently, the ooplasm is thoroughly mixed by microtubule-based streaming (reviewed in Mahajan-Miklos and Cooley 1994; Becalska and Gavis 2009).

Another important aspect of ovarian follicle development is specification of the anterior-posterior and dorsoventral polarity that is critical for the proper embryonic development (for review, see Bastock and St Johnston 2008; Roth and Lynch 2009). Within the ovarian follicle, nurse cells occupy invariably the anterior pole, while the oocyte resides at the posterior one. The establishment of the axes is aided by the reciprocal signal exchange between germline cells and the surrounding somatic (follicular) cells. Subsequently, the oocyte becomes polarized by asymmetric distribution of specific developmental factors in distinct regions of the ooplasm. The formation of a distinct pole (germ) plasm domain is a morphological manifestation of the early oocyte asymmetry.

3.1 Pole Plasm Contains Specific Ribonucleoprotein Complexes

Expression-based analyses have identified numerous factors localized to the pole plasm. The biochemical nature of many of these factors has been uncovered over the last three decades. It has been shown that a number of maternally provided proteins (e.g., Oskar, Vasa, Tudor, and Aubergine) and mRNAs (including oskar, cyclinB, nanos, polar granule component, and germ cell-less) are enriched in Drosophila pole plasm (Fig. 5.2a) and at least some of these factors are engaged in PGC specification, maintenance, and migration (Lecuyer et al. 2007; Frise et al. 2010; Gao and Arkov 2013; Jambor et al. 2015). In addition, certain pole plasm constituents (e.g., nanos mRNA) are necessary for the specification of the abdominal region of the future embryo (Wang and Lehmann 1991).

Fig. 5.2
figure 2

Pole plasm in Drosophila melanogaster. (a) Fragment of the oocyte (o) with distinct pole plasm (asterisk) labeled by anti-Vasa antibody. Whole-mount preparation. (b) Pole plasm containing polar granule precursors (arrowheads) labeled by anti-Vasa antibody. y Yolk sphere. Immunogold EM

Full size image

The pole plasm-specific mRNAs appear to share three important characteristics. First, they associate with proteins and are transported in the form of ribonucleoprotein (RNP) particles. Second, they are translationally repressed in the unlocalized form. Third, localized mRNAs are protected from premature degradation or decay until they fulfill their function/s (Gavis and Lehmann 1994; Kim-Ha et al. 1995; Markussen et al. 1995; Rongo et al. 1995; Smibert et al. 1996; Rangan et al. 2009). Despite these similarities, molecular analyses and live cell imaging revealed that various constituents of RNP particles are targeted to the oocyte posterior pole by distinct mechanisms. Four localization scenarios have been reported so far: (1) active transport along microtubules, as for oskar RNP particles (Clark et al. 2007; Zimyanin et al. 2008); (2) localized translation within the posterior ooplasm and repression elsewhere, as for Osk (Kim-Ha et al. 1995); (3) passive diffusion within ooplasm and local entrapment at the posterior region, as for cyclin B, germ cell-less, and nanos mRNAs (Forrest and Gavis 2003); and (4) degradation of mRNA in the ooplasm except for the posterior pole, as for nanos mRNA (Zaessinger et al. 2006).

3.2 oskar Is a Critical Player in Drosophila Pole Plasm Formation

Among the factors targeted to the oocyte posterior pole, the transport and localization of oskar mRNA are probably best characterized. oskar is transcribed in the nurse cell nuclei during early stages of oogenesis. In the nucleoplasm, cis-regulatory elements in oskar transcripts are recognized by trans-acting proteins, which trigger assembly of initial RNP complexes. This process starts with splicing of the first intron in the osk pre-mRNA. It has been demonstrated that the splicing-associated modifications are necessary for proper cytoplasmic localization of the oskar transcripts later during oogenesis (Le Hir et al. 2001; Hachet and Ephrussi 2004). Furthermore, recent analyses indicate that oskar mRNAs are not regulated as single molecules but form multimolecular units containing many copies of oskar transcripts (Chekulaeva et al. 2006; Little et al. 2015). Following transport to nurse cell cytoplasm, additional proteins join existing oskar RNP complexes forming relatively large RNP particles. Mutant analyses and in vitro assays identified polypyrimidine tract-binding (PTB) protein as a promoting factor in the assembly of these particles (Besse et al. 2009). Subsequently, oskar RNP particles are transported into the ooplasm via ring canals. Transport of oskar RNP particles within the nurse cell-oocyte syncytium requires a polarized cytoskeletal network and is mediated by RNA-binding proteins (e.g., Staufen), appropriate motor proteins, and cis elements in 3′ untranslated region (3′UTR) of oskar mRNAs (reviewed in Kugler and Lasko 2009). The transport occurs in two phases: nurse cell to oocyte (in early oogenesis) and ooplasmic, directed toward the posterior pole of the oocyte (in mid-oogenesis). Although both phases are microtubule-dependent, they are regulated differently and engage different motor proteins. Transport from nurse cells into the oocyte requires dynein (Bullock and Ish-Horowicz 2001; Clark et al. 2007). The interaction between oskar RNP particles and dynein is mediated by two dynein-associated proteins: Bicaudal D and Egalitarian (Clark et al. 2007). The second (ooplasmic) phase of the transport depends on a plus end-directed microtubule motor protein, the heavy chain of kinesin 1 (Palacios and St Johnston 2002). Surprisingly, direct observations of movements of GFP-tagged oskar mRNA in living oocytes revealed that oskar particles are transported (along microtubules) in all directions (Zimyanin et al. 2008). It appears, however, that a slight bias toward the posterior oocyte pole is sufficient for the adequate posterior accumulation of oskar transcripts.

Previous studies suggested that the pole plasm components become anchored in the posterior ooplasm by an actin-based mechanism. More recent high-resolution live imaging revealed, however, rather unexpectedly, that the pole plasm anchoring is dynamic and requires constant trafficking of RNP particles at the posterior oocyte cortex (Sinsimer et al. 2013). According to a current model, motility of RNP particles depends on cortical microtubules, kinesin and dynein motor proteins, as well as interplay between the microtubular and actin cytoskeletons.

During transport, oskar mRNA is translationally repressed by the RNA-binding proteins. One such protein is Bruno. This trans-acting factor binds to dedicated sequences, called Bruno response elements, in 3′UTR of oskar mRNA (Kim-Ha et al. 1995; Chekulaeva et al. 2006). According to current hypotheses, Bruno interacts with Cup protein, which in turn binds to and inactivates the eukaryotic translation initiation factor-4E (Nakamura et al. 2004; Chekulaeva et al. 2006; Besse et al. 2009). It has also been proposed that formation of densely packed RNP particles in which oskar mRNAs become inaccessible to the translational machinery is also required for translational repression (Besse et al. 2009).

Once oskar mRNAs reach the posterior ooplasm and are properly localized, the binding of activators, such as Orb, releases the repression and translation starts (Chang et al. 1999; Kim et al. 2015). However, this process is not straightforward. Alternative translation initiation from two in-frame start codons in oskar mRNA generates two isoforms of Oskar protein: Short and Long, each with distinct properties (Markussen et al. 1995; Breitwieser et al. 1996). While the Long Oskar is responsible for anchoring oskar mRNA and Short Oskar isoform within the posterior cortical region of the oocyte, the more abundant Short Oskar interacts with several germline-associated mRNAs (e.g., nanos, polar granule component, and germ cell-less) and proteins (e.g., Vasa) and is necessary for nucleating the pole plasm as well as the proper patterning of the posterior region of the embryo (Markussen et al. 1995; Breitwieser et al. 1996; Vanzo and Ephrussi 2002). After fulfilling its functions, Oskar is phosphorylated by two kinases (Par-1 and GSK-3/Shaggy), and subsequently Short Oskar is targeted for proteosomal-dependent degradation by SCF(-Slimb) ubiquitin ligase (Morais-de-Sá et al. 2013). Apparently, the same pathway is used to remove mislocalized Oskar.

3.3 The Assembly of the Drosophila Pole Plasm Occurs in a Stepwise Sequence

Formation of pole plasm is initiated by oskar mRNA trafficking to the oocyte posterior cytoplasm during mid-oogenesis, followed by synthesis of Oskar protein. Expression of oskar in ectopic locations has indicated that it can attract other factors (mRNAs and proteins) to new locations and organize a functional pole plasm capable of inducing pole cell formation (Ephrussi and Lehmann 1992; Smith et al. 1992). Moreover, overexpression of oskar showed that its level of expression directly influences the size of the pole plasm and positively correlates with the number of PGCs (Ephrussi and Lehmann 1992; Smith et al. 1992). The mechanism by which Oskar recruits other pole plasm components was until recently unknown. However, two recent analyses of the high-resolution crystal structure of Drosophila Oskar have revealed that the protein contains two domains: C-terminal OSK domain and N-terminal LOTUS domain. The first domain binds RNA, whereas the latter directly interacts with Vasa, a DEAD-box RNA helicase (Jeske et al. 2015; Yang et al. 2015). Because the OSK domain binds the 3′UTRs of oskar and nanos mRNA in vitro, it has been suggested that Oskar may regulate certain aspects of RNA metabolism, e.g., stability, translation, and/or localization, through direct interaction with 3′UTRs (Yang et al. 2015).

The initial formation of the pole plasm domain is followed by next wave of localization of specific mRNAs (including oskar, nanos, cyclin B, and polar granule component) to the posterior ooplasm during advanced stages of oogenesis (Dalby and Glover 1992; Nakamura et al. 1996; Forrest and Gavis 2003; Sinsimer et al. 2011). Until recently, the dynamics of pole plasm formation have remained a mystery. However, quantitative single-molecule imaging demonstrated that nanos, cyclin B, and polar granule component transcripts are initially translocated as RNP particles containing single-mRNA molecules but merge into larger heterogeneous granules specifically at the posterior pole of the oocyte (Little et al. 2015). This mechanism effectively separates the germ cell-destined transcripts from others and ensures their proper segregation into the forming pole cells.

Aubergine is another important factor in the Drosophila germline development. It plays a role in early oogenesis by regulating oskar RNA translation, but it has also been implicated in pole cell specification, consistent with its presence in the pole plasm (Wilson et al. 1996; Harris and Macdonald 2001; Becalska et al. 2011). Recently, the results of cross-linking immunoprecipitation experiments have indicated that Aubergine acts as adhesive trap mediating retention of germ plasm-specific mRNAs within the posterior ooplasm (Vourekas et al. 2016). Aubergine is a member of the evolutionarily conserved Piwi family proteins, which associate with piwi-interacting RNAs (piRNAs) and facilitate silencing of transposomes in the germline (reviewed in Mani and Juliano 2013). It has therefore been suggested that Aubergine association with mRNAs may be mediated by low-specificity pairing between the piRNAs and the target mRNA. In a recently proposed model, Aubergine may facilitate packaging of germ cell-destined mRNAs into polar granules (see below). However, it is still not entirely clear how the selection of these mRNAs may occur. It is worth adding that Aubergine also functions in early embryonic development (Barckmann et al. 2015).

3.4 Polar (Germ) Granules

Ultrastructural studies revealed that the Drosophila pole plasm contains distinct large (up to 500 nm in diameter) roughly spherical organelles called polar or germ granules (for review, see Mahowald 2001). They are composed of electron-dense granulo-fibrillar material, which is not surrounded by membrane. Another characteristic feature of Drosophila polar granules is that they are frequently accompanied by mitochondria and, in activated eggs, by polysomes. Following fertilization, polar granules are segregated to the forming pole cells (Lerit and Gavis 2011).

Biochemical and molecular analyses established that the germ granules are aggregates of RNP particles (reviewed in Voronina et al. 2011; Gao and Arkov 2013; Lehmann 2016). Their constituents are synthesized during oogenesis and accumulate in the pole plasm of the oocytes/eggs and early embryos. Localization of various RNAs and proteins to the germ granules coincides with nurse cell dumping and the cytoplasm streaming within the oocyte. Although a list of polar granule components is still not complete, it is well documented that these organelles contain proteins such as Oskar, Vasa, Tudor, and Aubergine (Fig. 5.2b) and a number of distinct RNAs, both coding and noncoding (Hay et al. 1988, 1990; Lasko and Ashburner 1990; Dalby and Glover 1992; Nakamura et al. 1996; Forrest and Gavis 2003; Thomson and Lasko 2004; Jones and Macdonald 2007; Anne 2010; Sinsimer et al. 2011). It has been also established that a critical step in polar granule formation is Oskar’s direct interaction with Vasa followed by recruitment of Tudor, Piwi, and Aubergine (Breitwieser et al. 1996; Harris and Macdonald 2001; Arkov et al. 2006; Thomson et al. 2008; Kirino et al. 2009, 2010; Nishida et al. 2009; Anne 2010; Fig. 5.3). It has been suggested that Tudor serves as a scaffold for germ granule assembly, and its interaction with other germ granule components is mediated by 11 specific protein motifs, called Tudor domains. These domains bind to proteins that contain methylated arginine and lysine residues (Thomson and Lasko 2005; Arkov et al. 2006; Kirino et al. 2009, 2010; Arkov and Ramos 2010; Liu et al. 2010).

Fig. 5.3
figure 3

Schematic representation of the distribution of molecular constituents in the pole plasm of Drosophila melanogaster. Aub Aubergine, Osk Oskar, RNPs ribonucleoprotein particles, Vas Vasa. See text for further details

Full size image

The inner architecture of the polar granules has been recently unveiled using super-resolution single-mRNA in situ fluorescence hybridization combined with structured illumination microscopy. This study revealed that the Drosophila polar granules are highly ordered entities. While proteins (Oskar, Vasa, Aubergine, and Tudor) are distributed evenly throughout the granules, mRNA molecules (nanos, polar granule component, and germ cell-less) form homotypic clusters and specifically concentrate either in the center or periphery of the granules (Trcek et al. 2015). Interestingly, such a spatial distribution is maintained during embryogenesis, indicating its functional significance. In addition, recent high-resolution analyses also indicate, contrary to previous speculations, that the Drosophila germ granules are heterogeneous aggregates (Little et al. 2015; Trcek et al. 2015).

Apart from the “core” proteins such as Oskar, Vasa, and Tudor, the Drosophila polar granules also contain processing body (P body) and ER-associated proteins (Thomson et al. 2008). The link between germ granules and P bodies is of particular interest. The latter are cytoplasmic domains implicated in RNA storage, repression of translation, and/or RNA degradation (reviewed in Voronina et al. 2011). The similarity of composition and function may indicate that polar granules are in fact P body-related structures, specific for the germline. It is worth adding, in this context, that canonical P bodies have been also identified in the Drosophila nurse cells (Liu and Gall 2007; Lim et al. 2009; Jaglarz et al. 2011).

Drosophila polar granules share many components with the pole plasm. It appears, however, that these subcellular domains are not entirely equivalent. Significantly, polar granules do not contain oskar mRNAs. In experiments in which oskar transcripts were ectopically targeted to polar granules, both Vasa expression and PGC formation were perturbed, suggesting that oskar mRNAs separation from the polar granules is essential for proper PGC development (Little et al. 2015).

4 Pole Plasm in “Non-Drosophila” Holometabolous Insects

Drosophila-like modes of germline specification have been found in several representatives of most major lineages of holometabolous insects including Diptera (Zissler and Sander 1982; Gutzeit 1985), Mecoptera (Ando 1973, this chapter), Coleoptera (Lynch et al. 2011), Neuroptera (this chapter), Megaloptera (Suzuki et al. 1981), Lepidoptera (Berg and Gassner 1978), and Hymenoptera (Meng 1968; Zissler and Sander 1982; Bilinski 1991; Klag and Bilinski 1993; Lynch et al. 2011). Genetic analyses have indicated that in certain dipterans (Anopheles, Aedes, and Culex, Juhn and James 2006; Juhn et al. 2008), coleopterans (Callosobruchus, Benton et al. 2016), and hymenopterans (Nasonia, Lynch et al. 2011), the induction of the pole plasm (and pole cells) relies, as in Drosophila, on oskar orthologs. It has been additionally shown that regulatory interactions upstream and downstream of Nasonia oskar are largely conserved with Drosophila, suggesting that these two regulatory networks are homologous and have a common origin (Lynch et al. 2011; Quan and Lynch 2016). On the other hand, the absence of oskar from the genomes of such insects as Bombyx mori (Xia et al. 2004), Apis mellifera (Wienstock et al. 2006), and Tribolium castaneum (Richards et al. 2006) clearly coincides with the lack of the pole plasm and pole cells in these species. All these results led to the suggestion that oskar represents a key factor in the evolution of the cytoplasmic inheritance in Holometabola (Lynch et al. 2011; Quan and Lynch 2016). It has also been speculated that oskar originated at the base of Holometabola and that this gene has been subsequently lost several times in independent holometabolous lineages (Lynch et al. 2011; Quan and Lynch 2016). Finally, Ewen-Campen and colleagues (2012) have demonstrated that oskar first arose much earlier, in the common ancestor of Holometabola and Orthoptera, and that its well-known role in germline specification is a derived condition of holometabolans (Ewen-Campen et al. 2012). To this evolutionary scenario, we may add one remark: the secondary loss of the cytoplasmic inheritance is apparently related not only to the loss of oskar gene per se but also to the disappearance of the nurse cells, where oskar is, as a rule, transcribed. Such a situation most probably happened during the evolution of groups characterized by the neopanoistic ovaries, such as Megaloptera: Corydalidae (dobsonflies) (Fig. 5.4c) and Siphonaptera (fleas) (Simiczyjew and Margas 2001) (Fig. 5.1).

Fig. 5.4
figure 4

The pole plasm (or its absence) in the oocytes of holometabolous insects. (a) Hymenoptera, Cosmoconus meridionator (early vitellogenesis); (b) Neuroptera, Osmylus fulvicephalus (postvitellogenesis; courtesy of Prof. Janusz Kubrakiewicz, Institute of Experimental Biology, Wroclaw University); (c) Megaloptera, Corydalidae, Corydalus peruvianus (early vitellogenesis); (d) Diptera, Rhagio sp. (early vitellogenesis); (e) Diptera, Rhagio sp. (postvitellogenesis). The pole plasm (arrows) is clearly visible at the posterior pole. Note lack of the pole plasm in Corydalus oocyte (c). fe Follicular epithelium, nc nurse cell, oo oocyte, egg envelope (open arrowheads). (a) semithin section stained with toluidine blue; (b–e) semithin sections stained with methylene blue

Full size image

Morphologically, the pole plasm of holometabolans is similar to that of Drosophila. It occupies the posterior oocyte pole and contains characteristic particles (termed germ or polar granules) or electron-dense accumulations not organized in particles, termed nuage (Zissler and Sander 1982). Molecular composition and functions of the nuage are discussed in Jaglarz et al. (2011), Voronina et al. (2011), and Kloc et al. (2014). Usually, the pole plasm is clearly recognizable even at the level of a light microscope, as it is virtually devoid of any reserve materials, i.e., yolk spheres and lipid droplets (Fig. 5.4a, b, d, e). Subsequent stages of the morphogenesis of the pole plasm and polar granules have been described only in some representatives of Diptera (e.g., Drosophila, Miastor) and Hymenoptera (e.g., Cosmoconus, Lissonota, Pimpla) (Meng 1968; Mahowald and Stoiber 1974; Klag and Bilinski 1993; Mahowald 2001). In this context, we have analyzed formation of the polar plasm in the scorpionfly, Panorpa communis, a representative of the old and plesiomorphic insect order Mecoptera.

The ovaries of Panorpa are meroistic polytrophic, and each ovarian follicle consists of an oocyte and three nurse cells (Ramamurty 1964; Bilinski et al. 1998; Ando 1973). At the onset of previtellogenesis, numerous roughly spherical nuage aggregations arise in the cytoplasm of the nurse cells (Bilinski et al. 1998). As oogenesis progresses, these aggregations are transferred to the ooplasm (Fig. 5.5a, asterisks) and accumulate next to the posterior oocyte pole forming a disk-shaped pole plasm (oosome) (Ando 1973; Fig. 5.5b, arrows). The latter contains irregular accumulations of nuage material (Fig. 5.5c, arrowheads) that do not constitute defined particles or polar granules. Above results and previously published data suggest that in all holometabolans, constituents of the polar plasm/polar granules (germline determinants) are synthesized in the cytoplasm of the nurse cells. Subsequently, these macromolecules are transported to the oocyte cytoplasm (ooplasm) and are localized to the posterior pole.

Fig. 5.5
figure 5

The pole plasm in oocytes of Panorpa communis (Mecoptera). (a) previtellogenic oocyte; note spherical aggregations of nuage material (asterisks). (b) posterior pole of late vitellogenic oocyte; note the pole plasm (arrows). (c) posterior pole of late vitellogenic oocyte, TEM. The pole plasm (delimited with red lines) contains several accumulations of nuage material (arrowheads). fe Follicular epithelium, gl glycogen particles, l lipid droplets, y yolk granules. (a, b) semithin sections stained with methylene blue; (c) ultrathin section

Full size image

5 Pole Plasm in Paraneoptera

The pole plasm and/or pole cells have been described only in three paraneopteran groups: Psocoptera (book lice), Thysanoptera (thrips), and Aphidoidea (aphids). As the processes found in these taxa are apparently different, we will discuss them separately.

5.1 Psocoptera

In Liposcelis divergens, the posteriorly located pole cells are specified relatively early, after formation of cellular blastoderm. Interestingly, the “preformed” pole plasm and polar granules have not been identified in fertilized eggs of this species (Goss 1952). Apparently, further studies or even reinvestigation of germline specification in this insect taxon is badly needed.

5.2 Thysanoptera

Classical histological studies have revealed that the pole plasm and polar granules are present at the posterior end of an unfertilized egg of Haplothrips verbasci (Heming 1979). This is quite astonishing because all thysanopterans are characterized by neopanoistic ovaries that most probably evolved by a secondary loss of nurse cells (Stys and Bilinski 1990; Büning 1993, 1994) (Fig. 5.1). In this context, we have followed the formation of the pole plasm in another representative of Thysanoptera, Taeniothrips sp., and found that the putative pole plasm of this species arises at the onset of vitellogenesis (Fig. 5.6a) and is clearly recognizable as oogenesis progresses (Fig. 5.6b). EM analysis has additionally indicated that the pole plasm is electron-transparent and does not contain either polar granules or electron-dense nuage-like material (not shown). The latter observation implies that germline specification should be reinvestigated also in this insect lineage.

Fig. 5.6
figure 6

The pseudo-pole plasm in oocytes of Taeniothrips sp. (Thysanoptera). (a) early vitellogenesis; (b) late vitellogenesis. Note that pseudo-pole plasm (arrows) is apparently more “transparent” than the rest of the ooplasm. fe Follicular epithelium. Semithin sections stained with methylene blue

Full size image

5.3 Aphidoidea

Recent molecular analyses have allowed for identification of a preformed pole plasm in the pea aphid, Acyrthosiphon pisum (Lin et al. 2014). It contains protein ApVas1 encoded by an ortholog of Drosophila gene vasa. On the other hand, oskar (or its ortholog) is not present in the pea aphid genome, indicating that the aphid germ plasm evolved independently and is not homologous to the germ plasm of holometabolans (Lin et al. 2014).

6 Putative Pole Plasm in Dermaptera

Although dermapterans are one of the basally branched insect lineages, they have simple polytrophic ovarioles in which ovarian follicles consist of two cells only: an oocyte and a single nurse cell (Tworzydlo and Bilinski 2008; Tworzydlo et al. 2010). Therefore, the question arises: whether (or not) developing dermapteran oocytes/egg cells contain the pole plasm? To answer this question, we have analyzed advanced stages of oogenesis in three dermapteran species (Forficula auricularia, Chaetospania borneensis, and Opisthocosmia silvestris). To our surprise, the posterior oocyte pole of all three species is occupied by a morphologically recognizable yolk-free cytoplasm (Fig. 5.7). Initially, this cytoplasm region has a dome-like shape (Fig. 5.7a); later it flattens and firmly adheres to plasma membrane surrounding the posterior pole (Fig. 5.7b). At the EM level, this putative pole plasm contains elongated mitochondria and numerous vesicles of the rough endoplasmic reticulum (RER vesicles) filled with fine granular material (Fig. 5.7c). Morphologically identical RER vesicles have also been found in the cytoplasm of the nurse cells (Fig. 5.7d). Histochemical tests have shown that the RER vesicles (both present in the nurse cells and in putative pole plasm) are AgNOR positive (Fig. 5.7d, e), suggesting that the RER vesicles are formed in the nurse cell cytoplasm and (subsequently) are transported to the posterior oocyte pole. The role of the dermapteran (putative) pole plasm and its RER vesicles remains elusive; its implication in the germline specification will require molecular analyses. In this context, we should stress only that the RER vesicles found in dermapteran oocytes are morphologically clearly different from bona fide polar granules, as the latter are never membrane-bound.

Fig. 5.7
figure 7

The pseudo-pole plasm in dermapteran oocytes. (a) Chaetospania borneensis (early vitellogenesis); (b–d) Opisthocosmia silvestris (late vitellogenesis); (e, f) Chaetospania borneensis (late vitellogenesis). The pseudo-pole plasm (arrows) is devoid of reserve materials and contains RER vesicles filled with fine granular material (v in c). Similar vesicles are also present in the nurse cell cytoplasm (v in d). The vesicles are strongly AgNOR positive (e, f—encircled). Gc Golgi complex, m mitochondria, nc nurse cell, oo oocyte, y yolk spheres. (a) semithin histocryl section stained with propidium iodide; (b) semithin histocryl section stained with methylene blue; (c, d) ultrathin sections; (e, f) semithin histocryl sections stained with AgNOR technique

Full size image

7 Conclusions

Despite relatively numerous studies, evolution of germline specification among insects is still debatable. There are at least two alternatives of phylogenetic interpretation of accumulated data.

  1. 1.

    The pole plasm and pole cells arose relatively early—at the base of Dermaptera and Phalloneoptera (Paraneoptera + Holometabola), i.e., simultaneously with the appearance of the nurse cells (Fig. 5.1, black box) or in the common ancestor of Paraneoptera + Holometabola (Fig. 5.1, diamond). Subsequently, the cytoplasmic inheritance has been lost independently in several insect lineages and replaced by the zygotic induction. Interestingly, all data clearly indicate that the putative loses have been much more frequent among Paraneoptera than within Holometabola. Such a situation is not surprising, as reversals to the pseudoprimitive condition (i.e., cytoplasmic inheritance → zygotic induction) are more likely to occur from a less advanced stage than from the more advanced one.

  2. 2.

    The pole plasm evolved much later—in the common ancestor of all holometabolans (Fig. 5.1, triangle). The morphologically recognizable “pole plasm” found in some paraneopteran groups, and in dermapterans, evolved independently.

In our opinion, accumulated data speak in favor of the second hypothesis. The following arguments are in support of this view:

  • Although the gene oskar evolved as early as in the common ancestor of Orthoptera and Phalloneoptera, it gained its new germline function much later—at the base of Holometabola (Ewen-Campen et al. 2012).

  • The germ plasm of aphids is not homologous to that of holometabolans; its specification does not rely on an oskar ortholog (Lin et al. 2014).

  • The putative pole plasm of such non-holometabolous insects as Thysanoptera and Dermaptera does not contain either nuage accumulations or bona fide germinal granules; therefore, it cannot be interpreted as genuine pole plasm. For such morphologically distinct region of the oocyte, we propose tentatively the term “pseudo-pole plasm.”