The UBAP2L ortholog PQN-59 contributes to stress granule assembly and development in C. elegans

When exposed to stressful conditions, eukaryotic cells respond by inducing the formation of cytoplasmic ribonucleoprotein complexes called stress granules. Stress granules are thought to have a protective function but their exact role is still unclear. Here we use C. elegans to study two proteins that have been shown to be important for stress granule assembly in human cells: PQN-59, the ortholog of human UBAP2L, and GTBP-1, the ortholog of the human G3BP1 and G3BP2 proteins. Both proteins fall into stress granules in the embryo and in the germline when C. elegans is exposed to stressful conditions. None of the two proteins is essential for the assembly of stress induced granules, but the granules formed in absence of PQN-59 or GTBP-1 are less numerous and dissolve faster than the ones formed in control embryos. Despite these differences, pqn-59 or gtbp-1 mutant embryos do not show a higher sensitivity to stress than control embryos. pqn-59 mutants display reduced progeny and a high percentage of embryonic lethality, phenotypes that are not dependent on stress exposure and that are not shared with gtbp-1 mutants. Our data indicate that both GTBP-1 and PQN-59 contribute to stress granule formation but that PQN-59 is, in addition, required for C. elegans development. Author summary The formation of so-called stress granules is an adaptive response that cells and organisms put into action to cope with changes in internal and environmental conditions and thus to survive to stressful conditions. Although it is generally thought that stress granule formation protects cells from stress-related damage, the exact role of stress granules in cells and organisms is not well understood. Moreover, the mechanisms governing stress granule assembly, and if and how the ability to form stress granules is important for C. elegans development is still unclear. Our work focuses on two conserved proteins, known to be involved in stress granule assembly in mammalian cells, and investigates their role in C. elegans embryos. We find that these proteins are important but not essential to assemble stress-induced granules in C. elegans. We moreover did not observe a different sensitivity to stress exposure between wild-type and mutant developing embryos, suggesting that at least in these conditions these proteins do not exert a protective role.


Introduction
270 Taken together our results suggest that PQN-59 and GTBP-1 do not help embryos to better resist to 271 exposure to heat. Our results also indicate that PQN-59 has additional roles in adult life and during 272 development that are independent of GTBP-1 and stress granule formation.

274
Discussion 276 Here we have studied the function of two conserved proteins, PQN-59, the ortholog of UBAP2L, and 277 GTBP-1, the ortholog of G3BP1/2 in assembly of stress granules in worm embryos and in worm 278 germlines.
279 Both PQN-59 and GTBP-1 are cytoplasmic proteins that condense into granules in response to stress 280 exposure. In Drosophila melanogaster, Lingerer/PQN-59 and Rasputin/GTBP-1 interact in Yeast 281 Two Hybrid assays (28). In human cells, G3BP-1 and UBAP2L coimmunoprecipitate and mutations 282 in UBAP2L that abolish the interaction with G3BP-1, are unable to rescue the stress granule assembly 283 defect of UBAP2L depletion (12,25). C. elegans GTBP-1 was isolated in pull down of PQN-59 from 284 embryos (unpublished), and we found that PQN-59 and GTBP-1 interact in a Yeast Two-Hybrid 285 assay, supporting the hypothesis that PQN-59 and GTBP-1 are in a complex in C. elegans. In contrast 286 with their human orthologs, the interaction of these two proteins or their presence is not essential for 287 the formation of stress-induced granules, as revealed by looking at GTBP-1 or PQN-59 and TIAR-1.
288 However, the assembly of stress-induced granules in the absence of one or the other is impaired, as 289 in this condition the granules appear less numerous, less defined in their shape, and show a faster 290 dissolution timing after stress relief. This suggests that the association between PQN-59 and GTBP-291 1 is not essential to assemble stress-induced granules but it is important to preserve stress granule 292 properties.
293 Deletion of the RGG domain of UBAP2L results in the abolishment of all interactions with stress 294 granule components and impairs stress granule assembly (12). Here we show that deletion of the 295 RGG domain in PQN-59 does not result in defects in the number of stress granules, suggesting that 296 this domain is dispensable for stress granule nucleation in the C. elegans embryo.
297 Single depletion of GTBP-1 and PQN-59 did not reduce the average number of TIAR-1 granules but 298 granule number was highly variable, suggesting that PQN-59 and GTBP-1 do contribute to proper 299 TIAR-1 granule formation. Consistent with a contribution, when PQN-59 was depleted in a gtbp-1 300 mutant, the number of TIAR-1 granules was reduced. This indicates that PQN-59 and GTBP-1 are 301 not strictly essential for TIAR-1 stress-induced granule assembly but they facilitate their formation.
302 On the opposite, depletion of TIAR-1 did not result in a significant defect in the number of GTBP-1 303 and PQN-59 granules, suggesting that TIAR-1 may act downstream in the process of stress granule 304 formation in C. elegans embryos. 305 Altogether, our data show that none of these proteins is required for stress induced granule assembly.
306 So, whereas in cultured human cells G3BPs and UBAP2L are important to form stress granules in 307 many stress conditions (9-12,25,32), in C. elegans, stress-induced granules can form in the absence 308 of GTBP-1, PQN-59, and in the absence of both suggesting that either an essential nucleator of stress 309 granules has still to be identified in this model or that the presence of disordered proteins is sufficient 310 to assemble stress induced granules in worms. This is reminiscent of work in intestinal progenitor 311 cells in Drosophila where canonical nucleators are not required for stress granule formation (33).
312 Depletion and mutation of PQN-59 result in additional phenotypes such as slow growth, reduced 313 progeny and embryonic lethality, all in absence of stress. These phenotypes were not observed in 314 gtbp-1 mutant or depleted animals. A recent paper has shown that the human orthologs, G3BP1/2 315 inhibit mTORC1 signaling by targeting mTORC1 to the lysosome (34). One possibility is that the 316 phenotypes of pqn-59 mutant embryos are dependent on GTBP-1. For example, an excess of free 317 GTBP-1 (not in complex with PQN-59) could be deleterious for worms and embryos. Co-depletion 318 of both PQN-59 and GTBP-1 resulted in a weak rescue of the embryonic lethality and the reduced 319 progeny phenotypes of pqn-59 mutants, indicating that these phenotypes may partially depend on an 320 excess of free GTBP-1. However, this weak rescue suggests that PQN-59 has additional important 321 functions in embryos and worms that do not depend on GTBP-1. These yet to be identified functions 322 could contribute to the regulation of the response to stress or be completely independent on the role 323 of PQN-59 in stress granule assembly. Additional studies will be required to understand the molecular 324 functions of PQN-59. 325 Stress granules have been proposed to protect cells from stress. We find that exposure to heat stress 326 kills to the same extent wild-type, pqn-59 or gtbp-1 mutant embryos. This indicates that during 327 embryonic development, the exposure to heat stress results in developmental failure, whether 328 embryos are able to form proper stress granules or not.

Strains
332 The C. elegans strains used in this work are listed in Table 1. Worms were maintained on NGM plates 333 seeded with OP50 bacteria, using standard methods (35). All the strains were grown at 20°C and 334 incubated at 20°C after dsRNAs injections.
335 Mutant strains were generated using CRISPR/Cas-9 technology, as described in (36). Single-guide 336 RNAs and repair templates, as well as PCR primers used to detect and sequence the mutations, are 337 listed in Table 2 and Table 3 respectively. The pqn-59 mutant strain (generated in the N2 background 338 and in the JH3199 (gtbp-1(ax2055[gtbp-1::GFP])IV) background) was generated by introducing a 339 frameshift mutation leading to the appearance of a premature STOP codon. The pqn-59ΔRGG strain 340 was generated excising the RGG-rich region (from aminoacid position 122 to aa 189), not altering 341 the reading frame.    Table 4). For all genes, the dsRNA was produced with the Promega Ribomax RNA 365 production system. dsRNA was injected in L4/young adult hermaphrodites which were incubated at 366 20°C. Germlines or embryos collected from injected hermaphrodites were analyzed 24 hours after 367 injection. 368