Germ plasm anchors at tight junctions in the early zebrafish embryo

The zebrafish germline is specified during early embryogenesis by inherited maternal RNAs and proteins collectively called germ plasm. Only the cells containing germ plasm will become part of the germline, whereas other cells will commit to somatic cell fates. Therefore, proper localization of germ plasm is key for germ cell specification and its removal is critical for the development of soma. The molecular mechanism underlying this process in vertebrates is largely unknown. Here we show that germ plasm localization in zebrafish is similar to Xenopus and amniotes but distinct from Drosophila. We identified non muscle myosin II (NMII) and tight junction (TJ) components as interaction candidates of Bucky ball (Buc), which is the germ plasm organizer in zebrafish. Remarkably, we also found that TJ protein ZO1 colocalizes with germ plasm and electron microscopy (EM) of zebrafish embryos uncovered TJ like structures at early cleavage furrows. In addition, injection of the TJ-receptor Claudin-d (Cldn-d) produced extra germ plasm aggregates. Our findings discover for the first time a role of TJs in germ plasm localization.


Introduction
For instance, a connection between the prominent Balbiani body (BB), also called 77 mitochondrial cloud, and the germ plasm was first noticed in Xenopus (Heasman et al., 1984). 78 Then, germ plasm gets anchored at the vegetal pole during oogenesis and after fertilization is 79 passively inherited during the cleavage period at the forming furrows of the most vegetal 80 blastomeres (Ressom & Dixon, 1988; Tristan Aguero, Susannah Kassmer, Ramiro Alberio, 81 Andrew Johnson, 2017). At the blastula stage, germ plasm positive cells internalize into the 82 embryo and then start their migratory journey until they reach the gonads. However, the 83 molecular structure tethering germ plasm to the vegetal pole during the cleavage period of 84 Xenopus embryogenesis is not known. 85 In zebrafish egg, germ plasm is also localized to the vegetal pole like in Xenopus, but this 86 similarity of its positioning changes at the end of oogenesis (Dosch, 2015; Moravec & Pelegri,87 2020; Raz, 2003). After fertilization, germ plasm streams together with cytoplasm during 88 'ooplasmic segregation' into the forming blastodisc at the animal pole (Welch & Pelegri, 2014). 89 Subsequently, germ plasm localizes to the first two cleavage furrows, anchoring in four points 90 at the four-cell stage (Olsen et al., 1997;Raz, 2003;Yoon et al., 1997). Indeed, maternal 91 mutants affecting the first embryonic cleavages also interfere with germ plasm recruitment 92 (Nair et al., 2013;Yabe et al., 2007). The first described cytoskeletal structure tethering germ 93 plasm in zebrafish was described as furrow-associated microtubule-array (FMA) (Jesuthasan, 94 1998; Pelegri et al., 1999). However, the FMA starts to disassemble after the third cleavage, 95 leaving the molecular identity of the cellular structure anchoring germ plasm after the eight-96 cell stage unresolved. The germ plasm organizers zebrafish Buc, Xenopus Velo1 and Drosophila short Oskar (sOsk) 127 share the remarkable ability to specify germ cells in zebrafish (Krishnakumar et al., 2018). To address whether this localization mechanism is also conserved between zebrafish, Xenopus 131 and Drosophila, we injected mRNA encoding GFP-fusions of these germ plasm organizers into 132 1-cell zebrafish embryos (Fig. 1A). At 2.5-3 hours post fertilization (hpf), we compared the 133 localization of the GFP-fusion proteins to the germ plasm using an antibody against the 134 endogenous Buc proteins, which is tightly associated with the germ plasm (Bontems et al., pole (Anne Ephrussi, 1992). We used this approach to address whether Buc is targeted by the 158 machinery that localizes Osk in the Drosophila embryo. We fused the buc ORF to GFP and a 159 bicoid-3'-UTR to direct its translation to the anterior pole of Drosophila embryos ( Fig. 2A). As 160 a positive control, we used short osk ORF fused to bicoid-3-UTR (sosk) (Tanaka & Nakamura, 161 2008). Indeed, immunolabelling of stage 4-5 fly embryos showed that Osk-GFP is anchored at 162 the anterior cortex of the embryos and around the anterior nuclei of ectopically induced PGCs 163 ( Fig. 2B, D). By contrast, Buc was neither localized to the embryo cortex nor did it form 164 perinuclear foci (Fig. 2E). Although the strength of the Buc-GFP signal increased during the 165 onset of cellularization (stage 5), the protein was not anchored to the anterior cortex but 166 distributed in a gradient originating at the anterior pole (Fig. 2C, E). Furthermore, although 167 Buc-GFP was expressed in the anterior pole of the embryos, no cells were formed that showed 168 morphological features of ectopic PGCs (Fig. 2C, E). Indeed, Vasa protein or pgc mRNA labeling 169 in these transgenic embryos confirmed that sOsk specified ectopic PGCs anteriorly, whereas 170 Buc-GFP did neither recruit Vasa protein or pgc mRNA to the anterior pole nor caused the 171 formation of ectopic PGC ( Supplementary Fig. 3). These results show that Buc is not 172 recognized by the localization machinery in Drosophila that anchors sOsk to the cortex, 173 suggesting that zebrafish and flies use different mechanisms for germ plasm localization. To analyze the localization of Buc in the germline of chicken embryos, we used whole-mount 186 immunohistochemistry. Double labelling with Buc and Cvh antibodies shows colocalization of 187 Buc with the germ plasm of cleavage stage I and stage II chicken embryos (Fig. 3A, B). 188 To confirm the positioning of Buc at cleavage furrows, we co-labeled embryos with the 189 membrane marker pan-Cadherin at stage II. Buc co-localizes with pan-Cadherin suggesting 190 that germ plasm colocalizes with the membrane in the chicken embryo ( Supplementary Fig.  191   4). Nonetheless, the more widespread distribution of pan-Cadherin implies that Buc localizes 192 to a restricted region of the plasma membrane. Taken together, these results indicate that 193 germ plasm localization to the membrane is evolutionary conserved between zebrafish and 194 chicken embryos. 195

196
The Buc localization signal is part of the conserved N-terminal BUVE motif 197 We have previously shown that the 3'-UTR of the buc mRNA is not required for its localization 198 (Bontems et al., 2009). Therefore, we analyzed its amino acid sequence to identify the protein 199 domain of Buc that is essential for the localization to the four germ plasm spots at 3 hpf. We 200 generated systematic deletions of Buc fused to GFP (schematically shown in Fig. 4A (Fig. 4A, B, Supplementary Fig. 5E, F), suggesting that aa 11-88 contains the residues 214 sufficient to target the protein to the germ plasm spots. To confirm that Buc does not contain 215 other motifves involved in localization, we generated a deletion of the isolated motif aa11-88 216 (Buc∆11-88) in full length Buc. This protein did not localize (Fig. 4A To investigate the importance of these two potential PLD domains in Buc, the colocalization 232 of deletion variants with the germ plasm was analyzed. Therefore, mRNA of deletion variants 233 of the BucLoc domain fused to mCherry were injected into 1-cell embryos and the 234 colocalization to germ plasm aggregates marked by Buc-GFP was examined at 3 hpf. As a 235 positive control we used the entire BucLoc domain (aa11-88) that shows colocalization with 236 the endogenous germ plasm (Fig. 5B, C, quantification in E). To narrow down the localization 237 motif further, the N-terminal 20 amino acids were removed. Indeed, Buc 31-88 showed a 238 slight reduction in germ plasm localization. Interestingly, when we deleted ten C-terminal 239 amino acids (Buc31-78), localization was restored to nearly wild-type frequency (Fig. 5D, E, 240 Supplementary Fig. 7B). By contrast, deleting four additional N-terminal amino acids (Buc35-241 78) almost completely abrogated localization. Nonetheless, the first PLD between aa25 and 242 aa30 does not seem to be necessary for localization. 243 To examine the role of the second PLD, we generated internal deletions in Buc31-78. When 244 we removed the second PLD (∆64-71), no fluorescence could be detected in the embryos 245 ( Supplementary Fig. 7C) suggesting that aa64-71 might affect protein stability or translation. 246 Therefore, we analyzed two variants with five amino acid deletions within the second PLD 247 domain. Strikingly, removing parts of the second PLD (Buc62-66 or Buc67-71) caused no 248 clear reduction in germ plasm localization (Fig. 5D, E, Supplementary Fig. 7D, E). In contrast, 249 when we kept the second PLD domain intact, but removed C-terminal sequences (Buc31-71), 250 the localization efficiency dropped to 15% (Fig. 5D As Buc forms clusters with the germ plasm in the proximity of the cleavage furrows, we aimed 257 to identify the cellular structure that is essential for its anchorage. One candidate is the 258 furrow-associated microtubule-array (FMA), which was shown to be involved in tethering 259 germ plasm in zebrafish (Jesuthasan, 1998;Pelegri et al., 1999). As our results show that 260 BucLoc domain is sufficient for the localization of Buc to the germ plasm foci, we used this 261 protein motif as a bait to identify cellular binding partners directly by co-immunoprecipitation 262 followed by mass spectrometry analysis. Embryos were injected at 1-cell stage with mRNA 263 encoding BucLoc-GFP, lysed at the stage of the formation of germ plasm foci and 264 immunoprecipitated using GFP-tag (Fig. 6A). Embryos injected with mRNA encoding GFP were 265 used as a negative control, and transgenic embryos for full length Buc-GFP were used to 266 control for mRNA overexpression (Fig. 6A, B). 267 In this analysis, we found 1817 protein candidates that potentially interact with full length Buc 268 and BucLoc but not with GFP. From those, 213 proteins were strongly enriched for BucLoc 269 interaction (Supplementary Table 1  where germ plasm is localized (Fig. 7E). In contrast, we did not find these structures at the 305 cleavage furrows, where germ plasm is not accumulated (Fig. 7F). This result shows for the 306 first time that early zebrafish embryos have TJ-like structures already at the 8-cell stage, which 307 are positive for ZO1 protein. Our data show that the localization machinery of germ plasm is different between vertebrates 365 and invertebrates. We identified that the N-terminal Buc (aa11-88), is necessary and 366 sufficient for the localization of the protein, the zebrafish germ plasm organizer. This region 367 of Buc contains two PLDs, however they do not play a role in germ plasm localization in 368 zebrafish. We found NMII as an interacting partner and colocalizing protein for Buc, 369 suggesting that germ plasm might be anchored to one of the cellular structures through NMII. 370 Indeed, we provide evidence that germ plasm localizes to TJs in early zebrafish embryos and 371 overexpression of TJ receptor protein Claudin-d induces ectopic germ plasm formation. 372

Evolutionary conservation of germ plasm anchorage among vertebrates 373
Invention of multicellularity requires cell adhesion and germ plasm for sexual reproduction. 374 With our finding, it will be possible to address whether germ plasm localization at TJs was 375 already used at the origin of Metazoa or whether it is a derived mechanism acquired during 376 vertebrate evolution. The isolated BucLoc motif does not show homologies to known protein 377 domains, which did allow to deduce its biochemical function. However, it proposes vertebrate 378 species, which might use a similar localization mechanism for Buc like the zebrafish. Indeed, 379 we found with this approach that the germ plasm in the early chicken embryo also localizes 380 to the cleavage furrow. As the BUVE-motif containing BucLoc shows similarity to Buc-like ORFs it would therefore be exciting to study the localization of TJ proteins in their embryos during 383 early cleavages. 384 Our data suggest that the tethering of germ plasm is conserved among vertebrates, as we 385 observed identical positioning in Xenopus, zebrafish and chicken, but not in Drosophila, 386 suggesting a similar anchorage mechanism in fish, frogs and chicken. In contrast, Drosophila 387 sOsk is not targeted by the vertebrate localization system. Therefore, our results show that 388 the localization machinery of germ plasm is different between vertebrates and invertebrates. 389 Despite the functional equivalence of Buc and Osk which is previously shown (Krishnakumar 390 et al., 2018), we here provide evidence that Buc and Osk proteins use different mechanisms 391 to localize germ plasm. Germ cell specification activity of these germ plasm nucleators looks 392 conserved, whereas the mechanism of their localization seems to adapt to the architecture of 393 the embryo. Therefore, different localization mechanisms are consistent with the different 394 shapes of early embryos. 395

Germ plasm localizes at TJs at early cleavage furrows in zebrafish embryos 396
Furrow-associated microtubule-array (FMA) is previously described as cytoskeletal structure 397 tethering germ plasm in zebrafish (Jesuthasan, 1998; Pelegri et al., 1999). However, the exact 398 cellular structure responsible for the anchorage of the germ plasm in zebrafish was still 399 unknown. It was previously shown that NMII is involved in various cellular structures (Liu et  of the predicted PLDs within the N-terminus when we identified the region between aa11 and 479 88 to be essential and sufficient for the localization of Buc to the 4 germ plasm spots at the 480 cleavage furrows (Fig. 4). The detailed mapping showed however, that none of these two 481 potential PLDs within this sequence are essential, but short stretches C-terminal of them (Fig.  482  5). This result does not exclude that the PLDs are involved, as the additional regions might be 483 required for the proper presentation of the PLDs. However, co-immunoprecipitation and 484 Mass spectrometry analyses showed that the domain between aa 11-88 interacts with about 485 213 peptides including myosin light chain and Cldn-d (Fig. 6). Nonetheless, 213 peptides is a 486 high number of interactions and includes probably a number of unspecific interactions. For 487 instance, we purified five of ten subunits of the RNA-exosome complex and probably only one 488 subunit directly interacts with Buc. We therefore prefer to interpret these results in a In conclusion, we found that vertebrates and invertebrates utilize different germ plasm 531 localization mechanisms, with evolutionary conservation between vertebrates. We 532 discovered that TJs anchor germ plasm during early zebrafish embryogenesis and that germ 533 plasm in zebrafish is anchored to the TJs and via Cldn-d receptor protein.

Microinjection of 534
cldn-d induced extra germ plasm spots. Currently we believe that NMII interaction with Buc 535 drives the localization of germ plasm into the TJs, as shown in the following model (Fig. 10). Previously synthesized capped RNA was diluted with 0.1M KCl and 0,05% phenol red (Sigma 545 Aldrich, Hannover). 2nl of RNA was injected into 1-cell stage embryos using PV820; WPI 546 injecting apparatus (Sarasota, USA). Injected embryos were incubated in E3 medium at 28°C 547 until they reached the developmental stage of the phenotype evaluation. 548

16-cell injection assay of Cldnd-ΔYV 549
To study whether non-functional Cldn-d has an influence on matured TJs, we conducted cldn-550 dΔYV injections in 16-cell embryos. In this assay, we injected the RNA directly into two cells 551 next to a germplasm localizing tight junction. As a control we used uninjected and cldn-d RNA 552 (1/5000 (Pflanz et al., 2015)). Anti-mouse and anti-rabbit antibodies coupled to Alexa 488, 568 566 or 647 were used as secondary antibodies (Invitrogen, 1/1000). Embryos were imbedded in 567 DPX to provide clearing and to protect from bleaching. 568 569

Biochemical methods 570
Co-immunoprecipitation (Co-IP) 571 CO-IP was performed to identify Buc protein interactome. Each sample was prepared from 572 500 deyolked high stage embryos after homoginization on ice in lysis buffer (10 mM Tris 573 (pH 7.5), 150 mM NaCl, 0.5 mM EDTA, 0.5 % NP-40, 1x complete protease inhibitor cocktail 574 (Roche, Mannheim)). The supernatant was subsequently used for the Co-IP using a GFP-575 binding protein coupled to magnetic beads (GFP-Trap_M; ChromoTek, Planegg-Martinsried) 576 following manufacturers instructions. After pulling down, the magnetic beads and their 577 bound proteins were either incubated with 2x SDS loading buffer for 5 min at 96 °C and 578 analyzed via SDS-PAGE and western blotting or sent for mass spectrometry (Core Facility of 579 Proteome Analysis, UMG, Goettingen), as described previously (Krishnakumar et al., 2018). 580

Selection criteria for specifically interacting proteins 581
In total, 3464 protein candidates interacted. From those, 1817 candidates were identified that 582 interacted with both Buc-GFP and BucLoc-GFP. We were not interested in every candidate 583 for interaction with Buc-GFP, as they might interact with any other region outside of BucLoc. 584 Therefore, we applied a set of criteria to identify significant interacting candidates with 585 BucLoc. First, any peptide below a background threshold of five in BucLoc-GFP was considered 586 as not significant and were sorted out. Furthermore, only proteins with counts in BucLoc-GFP 587 that were at least twice as high as in the negative control GFP were considered as significant. 588 To further reduce overexpression artefacts, enrichment in the positive control and in the 589 sample had to be within a magnitude of +/-4-fold. Applying these selection criteria, the 590 number of potential BucLoc interaction proteins could be restricted to 213 interaction 591 candidates (see the Supplementary for the full list of mass spectrometry candidates). 592

Immunohistochemistry 593
Embryos were fixed and stained as previously described (Riemer et al., 2015) with the 594 following antibody concentrations. 595