Microtubule Assembly and Pole Coalescence: Early Steps in C. elegans Oocyte Meiosis I Spindle Assembly

How oocytes assemble bipolar meiotic spindles in the absence of centrosomes as microtubule organizing centers remains poorly understood. We have used live cell imaging in C. elegans to investigate requirements for the nuclear lamina and for conserved regulators of microtubule dynamics during oocyte meiosis I spindle assembly, assessing these requirements with respect to recently identified spindle assembly steps. We show that the nuclear lamina is required for microtubule bundles to form a cage-like structure that appears shortly after oocyte nuclear envelope breakdown and surrounds the oocyte chromosomes, although bipolar spindles still assembled in its absence. Although two conserved regulators of microtubule nucleation, RAN-1 and γ-tubulin, are not required for bipolar spindle assembly, both contribute to normal levels of spindle-associated microtubules and spindle assembly dynamics. Finally, the XMAP215 ortholog ZYG-9 and the nearly identical minus-end directed kinesins KLP-15/16 are required for proper assembly of the early cage-like structure of microtubule bundles, and for early spindle pole foci to coalesce into a bipolar structure. Our results provide a framework for assigning molecular mechanisms to recently described steps in C. elegans oocyte meiosis I spindle assembly.

imaging in C. elegans to investigate requirements for the nuclear lamina and for 23 conserved regulators of microtubule dynamics during oocyte meiosis I spindle 24 assembly, assessing these requirements with respect to recently identified spindle 25 assembly steps. We show that the nuclear lamina is required for microtubule bundles to 26 form a cage-like structure that appears shortly after oocyte nuclear envelope breakdown 27 and surrounds the oocyte chromosomes, although bipolar spindles still assembled in its 28 absence. Although two conserved regulators of microtubule nucleation, RAN-1 and γ-29 tubulin, are not required for bipolar spindle assembly, both contribute to normal levels of 30 spindle-associated microtubules and spindle assembly dynamics. Finally, the XMAP215 31 ortholog ZYG-9 and the nearly identical minus-end directed kinesins KLP-15/16 are 32 required for proper assembly of the early cage-like structure of microtubule bundles, 33 and for early spindle pole foci to coalesce into a bipolar structure. Our results provide a 34 framework for assigning molecular mechanisms to recently described steps in C. 35 elegans oocyte meiosis I spindle assembly. The XMAP215 family member ZYG-9 is known to be required for oocyte meiotic spindle 305 assembly (Yang et al., 2003), but the nature of this requirement remains poorly 306 understood. To examine its role, we first assessed the dynamics of ZYG-9 localization 307 in comparison to microtubules and ASPM-1. We used CRISPR/Cas9 to generate an in 308 situ fusion of GFP to the endogenous zyg-9 locus and genetic crosses to introduce the 309 mCherry::H2B fusion ( Figure 6A and Supplemental Figure 15A). In contrast to 310 GFP::ASPM-1, we did not detect GFP::ZYG-9 in association with the microtubule cage; 311 rather it was initially more diffusely present near chromosomes and subsequently 312 became enriched at multiple pole foci and also was present more diffusely throughout 313 the spindle during pole coalescence. Upon the establishment of spindle bipolarity, 314 GFP::ZYG-9 was enriched at the poles but in contrast to GFP::ASPM-1, GFP::ZYG-9 315 also was detected between the poles. ZYG-9 binds the coiled-coil TACC ortholog TAC-316 1, and both promote microtubule stability during early embryonic mitosis in C. elegans 317 (Bellanger et al., 2007;Bellanger and Gonczy, 2003). We therefore used CRISPR/Cas9 318 to generate an in situ fusion of GFP to the endogenous tac-1 locus and observed 319 localization dynamics similar to GFP::ZYG-9 (Supplemental Figure 15B). To summarize, 320 ZYG-9 and its partner TAC-1 were both present throughout spindle assembly and both 321 appeared more restricted in distribution than microtubules but, as assembly progressed, 322 more broadly distributed than ASPM-1. 323

324
We next examined ZYG-9 requirements after RNAi knockdown in transgenic strains 325 expressing GFP or mCherry fusions to TBB-2, ASPM-1 and H2B and observed multiple 326 defects during meiosis I spindle assembly. First, when imaging spindles marked with 327 GFP::TBB-2 or GFP::ASPM-1, microtubule bundles assembled to form a peripheral 328 cage shortly after NEBD, but some of the microtubule bundles were not restricted to the 329 periphery and instead passed through the interior of the chromosome occupied space 330 ( Figure 6B and 6C, Supplemental Figures 13 and 14, and Supplemental Movies 4-9). 331 Subsequently, foci of GFP::TBB-2 and GFP::ASPM-1 were observed not only at the 332 periphery surrounding oocyte chromosomes, but also between some of the bivalent 333 chromosomes. Moreover, small spindle-like structures often appeared to form around 334 individual or small groups of bivalents, in contrast to the peripheral spindle foci that 335 coalesced to form a bipolar spindle in control oocytes. 336

337
The abnormal dynamics of spindle assembly observed after ZYG-9 knockdown were 338 accompanied by a significant increase in the level of oocyte spindle microtubules 339  Figure 8A). We also observed increased microtubule levels throughout 342 the oocyte cortex during meiosis I (Figure 7). These increases were surprising because 343 ZYG-9 is required for the stability of astral microtubules during early embryonic mitosis 344 (Bellanger et al., 2007;Bellanger and Gonczy, 2003), and ZYG-9 orthologs in other 345 species promote microtubule assembly (Akhmanova and Steinmetz, 2015), although in 346 some contexts they also promote microtubule instability (Shirasu-Hiza et al., 2003). 347

348
We also observed a striking lack of pole stability and extensive chromosome 349 segregation defects after ZYG-9 knockdown. In control oocytes, early small pole foci 350 stably associated with each other over time ( Figure 1B, Supplemental Figures 2 and 16, 351 Supplemental Movies 1 and 12). In contrast, after ZYG-9 knockdown, pole foci marked 352 by GFP::ASPM-1 fused and then often broke apart as meiosis I progressed ( Figure 6C, 353 Supplemental Figure 16 and Supplemental Movies 13-15). Consistent with a role in pole 354 stability, chromosomes sometimes segregated into three masses during anaphase (10 355 of 20 oocytes), although in other cases no segregation (3 of 20 oocytes) or segregation 356 into two masses (7 of 20 oocytes) were observed (Supplemental Figures 5C, 13 and  357 14). Finally, we observed similar defects throughout oocyte meiosis I after knocking 358 down the TAC-1 binding partner for ZYG-9 (Supplemental Movies 16 and 17), indicating 359 that ZYG-9 and TAC-1 have similar if not identical requirements. In summary, ZYG-9 360 and TAC-1 restrict cage microtubule bundles to the periphery and promote pole 361 coalescence and stability. Notably, they also appear to limit both spindle and cortical 362 microtubule levels during oocyte meiosis I. We have examined the requirements for factors involved in C. elegans oocyte meiotic 369 spindle assembly, with the goal of assessing their roles during a sequence of four 370 recently described assembly steps that generate these acentrosomal and yet bipolar 371 spindles during meiosis I. Our results show that the nuclear lamina, comprised of the 372 single C. elegans nuclear lamin LMN-1, is required for assembly of the underlying cage-373 like microtubule structure, although this structure is not required for bipolar spindle 374 assembly. While knockdown of either of two conserved regulators of microtubule 375 nucleation, the small GTPase RAN-1 and TBG-1/γ-tubulin, did not prevent bipolar 376 spindle assembly or chromosome segregation, microtubule levels after both 377 knockdowns were reduced and spindle assembly dynamics were altered. We also 378 identified two additional contributions to assembly of the early cage-like network of 379 microtubule bundles. After knockdown of the XMAP215 ortholog ZYG-9, or its binding 380 partner TAC-1, the microtubule bundles were no longer restricted to the periphery but in 381 some cases passed through the space occupied by oocyte chromosomes. As reported 382 previously, these microtubule bundles were reduced in prominence in mutants lacking 383 the nearly identical minus-end directed kinesins KLP-15 and -16. Finally, both ZYG-9 384 and KLP-15/16 were required for early spindle pole foci to coalesce into a bipolar 385 structure, but in distinct ways. ZYG-9 knockdown resulted in a lack of pole stability 386 during coalescence, while pole foci failed to coalesce in klp-15/16 mutant oocytes. Our 387 results, considered in more detail below, document requirements for microtubule 388 nucleation, cage assembly and pole coalescence, key steps in the assembly of 389 acentrosomal oocyte meiosis I spindles in C. elegans. production of oocyte nuclei that were smaller in diameter, and the microtubule cage also 406 was reduced in diameter. Moreover, microtubule levels declined to a minimum earlier 407 than was observed in control oocytes, and the time required to progress through 408 meiosis I was reduced. 409 410 TBG-1 knockdown also did not prevent assembly of the microtubule cage, and its 411 diameter was similar to those in control oocytes. However, rather than proceeding to 412 form a network of peripherally located small spindle pole foci, the cage collapsed into a 413 condensed ball of microtubule signal surrounding the chromosomes. Subsequently, 414 microtubules emerged from the collapsed structure, and the pole marker ASPM-1 415 appeared in multiple foci that coalesced to form a bipolar spindle of normal length that 416 segregated chromosomes to the same extent as in control oocytes. In spite of these 417 changes in spindle assembly dynamics, we did not detect any defects in chromosome 418 segregation by the end of meiosis I after knockdown of either RAN-1 or TBG-1, with one 419 exception after TBG-1 knockdown, consistent with previous reports indicating the lack of 420 an essential requirement for either of these regulators during oocyte meiosis. 421

422
While the more normal sequence of assembly events after RAN-1 knockdown and the 423 more substantially altered assembly dynamics after TBG-1 knockdown suggest that 424 these two regulators play distinct roles, in neither case have we been able to determine While we did not detect a requirement for either RAN-1 or TBG-1 in assembly of the 448 early microtubule cage, we did find that the nuclear lamina, which directly overlies this 449 structure, is required for its assembly. RNAi knockdown of the only C. elegans lamin 450 LMN-1 nearly eliminated the peripheral microtubule bundles. Restricting the assembly 451 of early microtubule bundles to the periphery to form a cage-like network might promote 452 pole coalescence by having it occur only along the inner surface of the nuclear lamina, 453 rather than throughout the volume occupied by oocyte chromosomes. Consistent with 454 such a role, the time required to complete meiosis I in control oocytes differs due to 455 variability in the time required for pole coalescence, and the reduced time required to 456 complete meiosis after RAN-1 knockdown occurs during pole coalescence and 457 correlates with a reduction in the diameter and surface area of the cage-like network. 458