Gradual recruitment and selective clearing generate germ plasm aggregates in the zebrafish embryo

Abstract

Determination of primordial germ cells (PGCs) is one of the earliest decisions in animal embryogenesis. In many species, PGCs are determined through maternally-inherited germ plasm ribonucleoparticles (RNPs). In zebrafish, these are transmitted during oogenesis as dispersed RNPs, which after fertilization multimerize and become recruited as large aggregates at furrows for the first and second cell cycles. Here, we show that the number of recruited germ plasm RNPs is halved every cell cycle. We also show that germ plasm RNPs are recruited during the third cell cycle, but only transiently. Our data support a mechanism in which systematic local gathering of germ plasm RNPs during cytokinesis and threshold-dependent clearing contribute to forming germ plasm aggregates with the highest RNP number and germ cell-inducing potential.

Introduction

One of the earliest cell fate decisions in animal embryos is the differentiation of primordial germ cells (PGCs) from somatic cells. Two main mechanisms by which PGCs become induced are epigenesis involving inductive cell-cell interactions (as observed in mammals) and preformation involving inheritance of maternally derived germ plasm, a specialized structure containing ribonucleoparticles (RNPs) (as observed in Drosophila, C. elegans, teleosts, amphibians, and chickens).¹ Phylogenetic analysis indicates preformation is more derived as it has evolved independently over multiple lineages.² Though the two mechanisms are different, many mRNAs are highly conserved in PGCs regardless of whether they are induced via epigenetic or preformative mechanisms. For example, the product for the gene vasa is expressed in germ plasm and/or germ cells in lineages from planaria to humans.³ Thus, germ plasm can be regarded as a specialization to achieve the effective accumulation of factors that function in a conserved PGC-determination program.

Strategies for the transmission of germ plasm components from one generation to the next typically involve two main stages, transmission into the egg during oogenesis, which often involves its localization in the oocyte, and incorporation of germ plasm to the PGCs during embryogenesis. In Drosophila, for example, transmission of germ plasm during oogenesis occurs through a complex multistep process involving a dynamic microtubule network and microtubule-based motor proteins as well as anchoring to the microfilament-based cortical network at the posterior pole of the egg. This so-called pole plasm then incorporates into cells that form prematurely at the posterior pole of the egg to generate PGCs.^4-6 In C. elegans, germ plasm-like granules, called P-granules, are detectable in the posterior region of the uncleaved zygote as prelocalized, cytoplasmic particles.^7,8 Under the influence of cell polarity PAR factors⁹ and the actin cytoskeleton, P-granules become restricted to the germ cell lineage during asymmetric cell division. In Xenopus, germ plasm components localize during oogenesis to the vegetal pole cortex of the egg after transport through microtubule dependent pathways, and after fertilization germ plasm granules coalesce into four large masses whose incorporation into vegetal cells confers PGC fate.¹⁰

In the zebrafish embryo, germ plasm components appear to utilize at least two different pathways.¹¹ Some germ plasm components, which include mRNAs for the genes vasa,^12,13 nanos¹⁴, and dead end,¹⁵ are found in the cortex of the animally located blastodisc, which forms the developing blastomeres. These “animal” germ plasm mRNAs become recruited to the furrows of the first and second cleavage cycles to form four large germ plasm masses. Other components, which include mRNAs for the genes deleted in azoospermia (dazl)^16,17 and bruno-like,^17,18 become localized during oogenesis to the vegetal pole of the oocyte in a process reminiscent of germ plasm localization in Xenopus.^19,20 Upon fertilization, zebrafish “vegetal” germ plasm mRNAs translocate animally along meridional cortical tracks and eventually become associated with animal germ plasm mRNAs that have already become recruited at the furrow.¹¹ At this stage, germ plasm forms four compact aggregates, even though these maintain a level of substructure with regards to the relative position of animal and vegetal germ plasm components.¹¹ As the embryo continues to cleave, at approximately the 32-cells stage, each of these masses ingresses into four PGCs, where they segregate asymmetrically during cell division until the activation of the zygotic germ cell program at the midblastula transition.^13,21

As the earliest observed major event in the pathway of germ plasm segregation in the zebrafish embryo, the mechanism by which animal germ plasm components become recruited to the furrows is of particular interest. Previous studies have shown that these germ plasm components appear to be present as single RNPs at the cortex of the newly forming blastodisc, which undergo increasingly higher levels of aggregation to generate four highly multimerized masses, each at the distal ends of the first and second cell cycle furrows.^11,22,23

Our previous work has described and defined stages of animal germ plasm RNP multimerization in the zebrafish embryo.^11,23,24 In a first stage, occurring prior to furrow initiation, singletons aggregate in an outward wave emanating from the center of the blastomere, generating multimers containing a small number (up to 17 observed in our studies) of germ plasm RNPs. We have termed this stage preaggregation to emphasize that RNP multimerization occurs prior to furrow formation.^11,23 During furrow initiation, in a stage of furrow recruitment, such aggregates become recruited to the forming furrow, generating larger rod-shaped multimerized aggregates in medial and distal regions of the furrow. As the furrow completes, in a stage of distal compaction, the rod-like RNP aggregate further compacts at the distal-most edge of its original range, forming a tight mass of germ plasm at each end of the furrow. The process is repeated for the second cell cycle, generating two more compact masses of germ plasm.

Previous studies^11,23 have begun to shed light on the underlying cellular mechanisms driving germ plasm RNP multimerization prior to and during furrow initiation, corresponding to the stages of preaggregation and furrow recruitment. During fertilization, a sperm-derived pair of centrioles reconstitutes, using maternally-derived components, a centrosome, which in turn nucleates a sperm aster prior to initiation of the first cell division cycle.^11,25,26 As sperm aster microtubules grow, they bind at their plus ends (those furthest away from the centrosome) germ plasm RNPs, and move them away from the center of the blastodisc.^11,23 This outward movement is coupled to germ plasm multimerization. Possibly the outward movement of RNPs generates local increases in RNP density that facilitate their aggregation, which may also be made possible by dynamic rearrangements of an actin network associated with the RNPs.^11,23 Outward growth of astral microtubules generates an “aggregation front” within the field of RNPs originally present at the cortex: at the inner rim of this field, which moves outwards due to microtubule growth, RNPs multimerize, while in outer regions RNPs have not yet undergone multimerization and remain as singletons.

A recent report by Nair et al.²³ has shown that the link between microtubule plus ends and germ plasm RNPs depends on the function of a duplicated copy of the birc5 or survivin gene, which in other systems codes for a component of the Chromosomal Passenger Complex known to be involved in various aspects of cell division including chromosome segregation and cytokinesis, as well as in cell survival.^27,28 The protein product of the zebrafish birc5b gene, also known as motley, is found associated with germ plasm RNPs at the blastodisc cortex.²³ In wild-type embryos, Birc5b protein is also associated with microtubule plus ends, consistent with a function for Birc5b as a linker between microtubules and germ plasm RNPs. Indeed, embryos mutant for motley exhibit defects in germ plasm RNP multimerization at the aggregation front, as well as defects in its recruitment to the furrow. In motley mutants, the mutated Bir5b protein appears to remain bound to germ plasm RNPs, yet these are not in contact with astral microtubule ends, emphasizing the importance of Birc5b-mediated microtubule-RNP linkage in the aggregation process.

While this model begins to explain multimerization corresponding to preaggregation, how is recruitment of RNPs to the furrow achieved? This process may be explained by a simple extension of this model, accommodating for the behavior of the spindle apparatus, specifically asters during cell division. While formation of the sperm aster, a monoastral structure, occurs prior to the initiation of the first cell cycle, cell cycles corresponding to blastomere divisions contain microtubule asters arranged in bipolar structures emanating from centrosomes at the spindle poles.^25,26 These asters reach the cortex and are thought to provide signals, including the Chromosomal Passenger Complex, to initiate furrow formation along the length of the large blastomeres present in the early zebrafish embryo,²⁹ and furrow initiation appears to occur at a microtubule-free zone that forms at the region of overlap between asters from opposite sides of the furrow.^23,29,30 Because these asters also grow during each of the early cell cycles, their outward movement allows them to continue the outward extension of the RNP aggregation front, contributing to RNP multimerization. Relevant to furrow RNP recruitment, the bipolar nature of the asters allows them to deposit germ plasm RNP aggregates from both sides of the furrows into the furrow itself. Indeed, rows of aggregates at the tips of astral microtubules can be observed immediately flanking the initiating furrows, presumably as they are being deposited by outward forces from each of the spindle poles.²³

These observations suggest a simple model where the processes of preaggregation and furrow recruitment have the same mechanistic basis: in both cases RNP multimerization is facilitated by microtubule growth from a radial source, a monoaster prior to the first cell cycle, and bipolar asters during cell division cycles. (The process of distal RNP compaction, on the other hand, no longer has a radial movement component, and instead exhibits a medial to distal translocation along the plane of the furrow). Due to the arrangement of microtubules, the fertilization monoaster can only contribute to preaggregation, while bipolar astral microtubules in subsequent cell cycles contribute to both preaggregation and furrow recruitment. Recruitment to the furrows, a landmark of germ plasm RNP segregation, therefore appears to represent a local gathering of RNP aggregates in quadrants immediately adjacent to the forming furrows. This model also indicates how germ plasm RNP recruitment is coupled to cell division, since the same dynamic cytoskeletal structure, growing astral microtubules, in addition to facilitating germ plasm RNP multimerization, brings both germ plasm RNPs and furrow initiating signals to the site of furrow formation.

Here, we probe further this model of RNP furrow recruitment in the early zebrafish embryo, involving the coordination of RNP multimerization to cytokinesis during blastomere division. We first show that, as predicted by a local gathering model, the number of RNPs that become recruited to the furrow decreases by approximately half each cell cycle. We further document that RNP recruitment includes, in addition to the furrows of the first and second cell cycles, the site of furrow formation for the third cell cycle. This indicates that RNP multimerization coupling to cytokinesis is a general process of the early embryo not restricted to the first two cell cycles, and raises the question of how specificity is achieved so that only four germ plasm aggregates segregate to PGCs as is observed. To examine this question, we determined the localization of germ plasm RNPs after the first two cell cycles and show that RNPs recruited to third cell cycle furrows are no longer detected during subsequent cell cycles. Together, our data provide support for a mechanistic model where germ plasm furrow recruitment is driven by symmetrical, radial sources participating in both furrow induction and local germ plasm RNP multimerization, and that the number of lasting germ plasm masses, and therefore induced PGCs, is determined by selective RNP aggregate stabilization.

Results

In order to visualize and quantify germ plasm RNPs as they become recruited to furrows corresponding to the first embryonic cell cycles, fixed embryos were labeled with antibodies against human serine-19 phosphorylated nonmuscle myosin II (p-NMII), a robust marker of germ plasm RNPs during furrow recruitment,²³ and α-tubulin, which allowed us to discern the stage within the cell cycle and the location of the furrow. Embryos were fixed at time points corresponding to furrow formation during the first through fifth cleavage cycles (Fig. 1 and data not shown). As expected, we detected germ plasm RNP accumulation at first and second cell cycle furrows (Fig. 1A, B, D, and E). We also observed significant RNP accumulation to third cell cycle furrows, which has not been previously reported (Fig G,H). During furrow maturation at these stages, including the third cell cycle, RNP aggregates underwent the tightening and accumulation at the distal end of the furrow characteristic of distal compaction (Fig. 1C, F, and I). RNP accumulation in furrows corresponding to the fourth and fifth cell cycle was not detected, although during these two later stages the anti-myosin II antibodies ceases to label germ plasm RNPs.²³

Figure 1. Embryos undergo germ plasm RNP recruitment to furrows during the first three cleavage cycles. WT embryos fixed during furrow initiation for the first (A), second (D) and third (G) cleavage cycles. Recruitment was measured by averaging counts of RNPs found in each furrow for the 1st (B), 2nd (E), and 3rd (H) cell cycles during furrow initiation. RNP distal compaction was observed in mature furrows for the 1st (C), 2nd (F), and 3rd (I) cleavage cycles. Anti-p-NMII (green in A,D,G, grayscale in B,C,E,F,H,I) labels RNPs and anti-α-tubulin (red in A,D,G) labels microtubules. Panels in grayscale correspond to representative high magnification images corresponding to furrow initiation (B,E,H, to show RNP furrow recruitment, at stages similar to those in overviews in A,D,G, respectively) and furrow compaction (C,F,I, to show distal compaction at a slightly later stage in furrow maturation during the same cell cycle). Scale bars: 100 μM (A,D,G) and 10 μM (B,C,E,F,H,I)

We quantified RNP aggregates observed in furrows for the first through third cell cycle by manually counting RNPs in stacks of confocal images of the various furrows. We found that the average number of RNPs is relatively constant for each furrow, and that the number recruited at each furrow is approximately half as many RNPs as that at the furrow for the previous cycle (Fig. 1B, E, and H). Thus, the average number of RNPs at the furrow for the first cell cycle was 1,162 +/− 355, for the second cell cycle 608 +/− 104, and for the third cell cycle 240 +/− 70. After furrow maturation, this results in compact germ plasm masses of decreasing sizes (Fig. 1C, F, and I). The approximate halving of the number of RNPs for each subsequent cleavage cycle is consistent with the fact that the cortical area corresponding to a dividing blastomere is halved each cell cycle. The pattern of RNP recruitment reduction through time suggests that, at least for the first three cell cycles, furrow RNP recruitment during a given cell cycle is relatively independent of that in other cell cycles, as would be the case if gathering of RNPs at the furrow is local, encompassing only cortical regions closest to the newly forming furrow. Since cleavage planes for subsequent cell cycles are perpendicular to the previous one,^31,32 local recruitment of RNPs coupled to an alternating cleavage orientation pattern provides a mechanism for the effective gradual gathering of most cortical RNPs in the cleaving embryo.

The observation that germ plasm RNPs become recruited to furrows corresponding to the first three cell cycles contrasts with previous findings that only four masses of germ plasm become integrated into each of four PGCs.^11-13,21 This suggested that only masses at first and second cell cycle furrows become stabilized, and that masses that may become recruited in furrows for subsequent cell cycles, such as the third furrow, become degraded. We tested this hypothesis by visualizing germ plasm components during the first through fifth cell cycles, this time using vasa RNA as a germ plasm RNP marker that is maintained throughout these stages.^11-13,21 We find that although forming furrows recruit RNP aggregates during the third cell cycle (Fig. 2A), third furrow masses are undetected or barely detected at later stages of development, such as when the fourth cell cycle is completed (Fig. 2B) or during the fifth cell cycle (data not shown). In addition, while a significant number of RNPs can be observed in cortical, non-furrow regions up to the third cell cycle, this number appears to be reduced during the fourth and subsequent cell cycles. These observations are consistent with a model where only germ plasm masses that become recruited to the first and second cell cycle furrows are stable, and other cortical germ plasm RNPs, including masses that form at the third cell cycle furrows, become degraded. Because of the observed pattern of reduction in RNP furrow recruitment during blastomere cleavage, a simple hypothesis that fits the data is that a threshold amount of RNPs must be present in germ plasm aggregates for their stable maintenance past the third cell cycle.

Figure 2.vasa, a germ plasm component, is recruited to third cell cycle furrows; however vasa is not maintained in third cell cycle furrows after the fourth cell cycle. Wild-type embryos fixed and hybridized using probes against vasa RNA during the third (A) and fourth (B) cell cycles. RNP furrow recruitment and distal compaction occurs in the furrow corresponding to the first three cell cycles (A); however during the fourth cell cycle the four aggregates recruited to third cell cycle furrows are reduced (as shown in insert “3”) or undetectable (B). Inserts labeled “1,” “2,” “3” indicate higher magnification images of RNP aggregates recruited to the furrows for the first, second, and third cell cycle, respectively, and insert labeled “c” shows a representative cortical (non-furrow) region. Scale bars: 100 μM (A, B), 10 μM (insets).

Discussion

Individuals in any population must develop a robust mechanism to induce PGCs, essential for reproduction. Zebrafish embryos employ a preformative mechanism using maternally derived germ plasm RNPs. We have previously proposed a mechanistic model for the gathering of these RNPs during the first cell cycles of the early embryo.^11,23,24 This model relies on the outward movement of RNPs caused by growing astral microtubules, which facilitate by RNP multimerization prior to furrow initiation (preaggregation). While this process begins immediately after fertilization, when microtubules emanate from the monopolar sperm aster (Fig. 3A,B), it continues during the first several mitotic cycles when asters emanate from a bipolar spindle (Fig. 3C-E). Radially outward movement of RNP aggregates from opposite sides of the spindle leads to the gathering of RNPs at the forming furrow (furrow aggregation). These early cleavages are highly regular and alternate their orientation by 90° in each subsequent cell cycle.^31,32 In this model, the alternating cleavage pattern, coupled to the symmetrical gathering at the furrow from asters in opposite sides of the cleavage plane, results in the gradual gathering of cortical RNPs to the furrows.

Figure 3. Systematic, gradual germ plasm RNP recruitment to the furrows is followed by the stabilization of germ plasm aggregates in furrows for the first two cell cycles. Microtubules emanate from the sperm monoaster during the first cell cycle (A) reaching the cell cortex where they initiate preaggregation of germ plasm RNPs in an outwardly moving “aggregation front” (B). During cytokinesis for the first (C), second (D), and third (E) cell cycles, RNPs dispersed in the blastodisc cortex are gradually collected by the action of asters from opposite sides of bipolar spindles. The orientation of cell division plane alternates by 90° every cell cycle (double-headed arrow indicates direction of the spindle for each ongoing cell division, which is perpendicular to the plane of division), allowing the collection of RNPs from cortical regions not harvested by the previous cell cycle. Recruited RNPs undergo distal compaction during furrow maturation, resulting in a more compact structure at distal ends of the furrow. RNP clearing results in the depletion of the four RNP aggregates generated during the third cell cycle and any remaining uncollected cortical RNPs, so that subsequent cell cycles (F) contain only the largest RNP aggregates, corresponding to the first and second cell cycle furrows. Furrows for the first (1st), second (2nd), third (3rd) and fourth (4th) cell cycles are indicated (red font indicates actively forming furrow). Furrow maturation generates a mature furrow septum containing newly added membrane and cell adhesion junction components.

Here, we provide further data supporting this model. We show that with each subsequent cell cycle the number of germ plasm RNPs that become recruited to the furrow is approximately halved. This is consistent with the expectation that the cortical surface decreases by half during each cell cycle due to cell division, and a mechanism where microtubule asters facilitate the local gathering of RNPs in quadrants immediately adjacent to the furrows.

We also show that, although for over a decade germ plasm components have been reported to localize to the furrows of only the first and second cell cycle, there is significant RNP recruitment to furrows for the third cell cycle. This is also consistent with our mechanistic model, which predicts that the process of germ plasm recruitment might continue during cell division for as long as there are germ plasm RNPs remaining at the cortex and as long as the recruitment machinery is in place. In terms of this machinery, there is no evidence indicating that the first two cell cycles are different from the third and subsequent cycles. Although the earliest cleavage cycles in the zebrafish embryo show slow calcium waves associated with the furrows, and such waves might provide a differentiating property to these furrows, these calcium waves are still present in third cell cycle furrows.³³ Other components are known to function with temporal differences in early embryos of the zebrafish and other species, although such changes have been found to be gradual and to occur at later stages, during the mid- to late- cleavage.^23,34 Thus, there is no reason to propose that furrows for the third cycle might have different cellular machinery from those for the first two cycles. Indeed, recruitment of a predictable number of RNPs to the third cell cycle furrow suggests a common mechanism for RNP recruitment during the first several cycles, including the third. Instead, germ plasm RNP recruitment may cease when cortical RNPs become exhausted through (1) recruitment during the alternate pattern of cell cleavage, (2) outward movement of the aggregation front as asters reach peripheral areas of the cortex, and (3) degradation of germ plasm RNPs not assembled as large aggregates (see below).

Even though recruitment and distal compaction occur in the third cell cycle furrow, generating eight germ plasm RNP masses, ample evidence indicates that the zebrafish blastula embryo contains only four germ plasm masses corresponding to those for the first and second cell cycles, each of which later becomes incorporated into each of four PGCs.^11-13,21 This suggests that the additional four masses of RNPs recruited to third cell cycle furrows are not maintained in subsequent cell cycles, unlike the initial four RNP masses recruited to furrows for the first two cell cycles. Given the trend in size of recruited germ plasm RNP aggregates, which diminishes each cell cycle, one simple explanation may be that there is a threshold amount of RNPs that must be met for maintenance of germ plasm masses, and that aggregates that fall below this threshold become degraded (Fig. 3F). Indeed, we detect beginning at the fourth cell cycle a general decrease in the number of cortical RNPs that coincides with the disappearance of RNP masses from third cell cycle furrows.

Together, our findings support a model for germ plasm RNP furrow recruitment in the early zebrafish embryo dependent on symmetric gathering of RNPs mediated by astral microtubules emanating from the poles of bipolar spindles. This model includes several general mechanistic principles.

Cooption of cellular machinery

In the zebrafish embryo, we find common elements in furrow induction and germ plasm RNP recruitment to the furrow. In both cases, growing astral microtubules bring furrow-inducing factors, such as the CPC, as well as germ plasm RNPs to the furrow. Moreover, CPC components are involved not only in furrow induction but also in germ plasm multimerization even prior to furrow formation. This cooption of cytokinesis machinery for RNP recruitment may also occur in other organisms where germ plasm gathers at the site of furrow cleavage, such as in medaka³⁵ and chick³⁶ embryos. This cooption may have evolved as a simple mechanism to generate a small number of large germ plasm masses in different quadrants of the embryo, which would allow for the effective gathering immediately after fertilization of dispersed germ plasm components. Recent studies have shown a role for germ plasm factors in pluripotency,^3,37 and have shown an association of pluripotency factors with the remnants of the microtubule cytoskeleton during cytokinesis, the midbody.³⁸ Thus, it is also possible that the observed common mechanism of RNP segregation and furrow inducing factors reflects ancient links between these two processes.

Systematic gathering of RNPs through space

Local recruitment of dispersed RNPs during cytokinesis, combined with cleavage division planes alternating 90° every cell cycle, may constitute an effective mechanism to gather a large fraction of germ plasm RNPs into large masses. There is some inherent loss of RNPs in this system, namely granules that are pushed toward the periphery of the blastodisc but that do not become recruited to any of the furrows. However, a large fraction of granules, those closer to forming furrows during the first three cycles, do undergo furrow recruitment. Characterizing cellular parameters influencing the effectiveness of this process awaits the developing of imaging tools to track RNPs in the blastodisc and cleaving embryo.

Multiple inputs for temporal selectivity

Although the output of recruitment involves large aggregates solely in the furrows for the first and second cell cycles, the underlying furrow recruitment mechanism is not selective for these cycles alone: the embryo is competent to recruit germ plasm RNPs also during at least the third cell cycle. Temporal selectivity of the response is therefore not intrinsic to the RNP gathering system but is instead provided by a combination of inputs involving a subsequent step of RNP clearing. The trend of decreasing RNP aggregate size inherent in the RNP gathering system suggests that RNP clearing may be guided by aggregate threshold size, although we have not ruled out other possibilities for differential aggregate stabilization. Regardless of the precise mechanism, only larger RNP masses, those corresponding to the first two cell cycles, become stabilized. Thus, a specific response, furrow recruitment in the first and second cell cycles, is achieved through a general mechanism not temporally restricted to the stages exhibiting the response.

It is likely that the goal of germ plasm inheritance is to allow cells to more readily steer toward the PGC determination pathway. Zebrafish embryos appear to achieve this goal by systematically gathering RNPs to generate germ plasm masses and subsequently stabilizing only those with the greatest number of germ plasm RNPs, i.e., those with the greatest potential to induce the PGC fate. These studies begin to form a picture of germ plasm RNP furrow recruitment in the context of the early zebrafish embryo subcellular architecture. Future studies will provide further mechanistic detail in this process, as well as in other aspects of germ plasm segregation in this organism, such as distal compaction during furrow maturation and germ plasm asymmetric segregation during cell division in the mid-cleavage stages.

Materials and Methods

All animals were handled in strict accordance with good animal practice as defined by the relevant national and/or local animal welfare bodies, and all animal work was approved by the appropriate committee (University of Wisconsin-Madison assurance number A3368–01).

Fish stocks were raised and maintained under standard conditions at 28.5°.³⁹ Wild-type fish were mated and embryos collected within 5 min. Embryos were dechorionated and fixed using microtubule fix containing glutaraldehyde and post-fixed with methanol as described²² at time points corresponding to the first five cleavage cycles (45 min post-fertilization (mpf), 60 mpf, 75 mpf, 90 mpf, and 105 mpf). Antibodies against p-NMII and α-tubulin were as described.²³ Embryos were imaged using a Zeiss 510 confocal LSM and images analyzed using FIJI.⁴⁰

RNP counts within aggregates at furrows were made manually while scanning a stack of confocal images, marking individual RNPs to avoid counting RNPs more than once. Counting germ plasm RNPs manually was found to be more accurate than using automations based on labeled surface area, which led to gross over-estimation for highly aggregated RNPs possibly due to a disproportionate increase in overall labeling intensity due to RNP signal overlap. Seven furrows were counted for the first and second cell cleavage each, and ten furrows were counted for the third cleavage. Anti-tubulin labeling allowed determination of the precise location of furrows, which contains a sharp microtubule-free zone.²³

For florescent in situ hybridization, embryos were synchronized and fixed at the same time points as above. Fixed embryos were labeled using a digoxigenin-labeled antisense RNA probe for the gene vasa,¹² detected using Alexa-488 Tyramid as described,^11,41 and imaged using a Zeiss 510 confocal LSM. Images were stitched using FIJI plugin for 3D stitching by S Priebisch.

http:dx.doi.org/10.4161/bioa.26538