Zhang Chujun

Activation of IRE1 and PERK: Amino acid starvation during the KRB incubation leads to the activation of the UPR (unfolded protein response). In Chapter 3, we show that KRB elicits ER stress and activation of two downstream kinases, IRE1 and PERK leading to the stimulation of UPR. probably occur through the clustering of two ER membrane kinases IRE1 and PERK. Interestingly, UPR stimulation has been shown that linked to the MARylation Drosophila PARP16 (Jwa and Chang, 2012), which is also a transmembrane protein of the ER and the activation of PARP16 in KRB is essential for Sec body formation (Aguilera-Gomez et al., 2016). Whether the UPR stimulation during KRB incubation induces Sec body formation via this mechanism or independently remains to be further investigation. It could also be events that are downstream of the UPR, the Bip upregulation itself, and/or cytoplasmic changes. One interesting aspect is whether the activation of UPR might affect biophysical properties, membrane association dynamics, and conformation and post-translational modifications of Sec16 and COPII subunits, which would lead to their enhanced phase separation properties under stress. Importantly the sole stimulation of the UPR alone (through DTT) is not sufficient to form Sec bodies.

KRB appears to be a combination of UPR stimulation through amino acid starvation and SIK stimulation by a moderate salt stress. To form Sec bodies, the stimulation of both IRE1 and PERK needs to be combined to a moderate NaCl stress, through the activation of SIKs. Interestingly, a high NaCl stress, even in the presence of amino-acids also induces the formation of Sec bodies in a sufficient manner. It therefore suggests that the absence of amino acid potentiates the moderate NaCl stress. At present, this is not understood, as IRE1 and PERK inhibitors do not influence SCHNSO-induced Sec body formation, so the exact link between IRE1, PERK, SIKs and the UPR remains to be further investigated. Alternatively, the salt stress could potentiate the UPR kinases but it does not appear to be via SIKs, as HG-9-91-01 (the pan-SIK inhibitor) does not modulate the UPR, yet strongly reduces Sec body formation.

The formation of Sec body and stress granules is governed by different intracellular pathways
As mentioned above, both KRB and SCHNSO (high NaCl stress) leads to the formation of Sec bodies as well as Stress granules. Both stress assemblies form by phase separation in the same time frame and in close proximity to one another, but without intermixing.

Stimulation of the UPR. Sec body formation is triggered by UPR activation combined to SIK activation, mimicked in the dish by addition of DTT and 100 mM NaCl. Inhibition of IRE1, PERK and SIK affects the KRB and SCHNSO induced Sec body formation, but interestingly not at all the stress granule formation. This is remarkable as it has been widely reported that stress granules form in response to ER stress (Kimball, 2003; Lin et al., 2007). It is also the case for S2 cells where DTT, Heat stress, APS leads to the formation of stress granules (Chapter 4). This strongly suggested that ER stress of S2 cells could be instrumental to stress granule formation, at least upon KRB treatment, as we have shown that it does stimulate ER stress (Chapter 3) (Zhang et al., 2021). However, none of the ER stress inhibitors (AMG-18, 4u8C, PERKi) or depletions (PERK and ATFS depletion) lead to a decrease in stress granule formation upon KRB and SCHNSO. This suggests that even though ER stress can lead to stress granule formation, the downstream kinases are not involved in stress granule formation in KRB and SCHNSO.

Taken together, these results suggest that the same cellular stresses induce the formation of two stress assemblies upon the activation of many pathways, a subset of them specifically leading the formation of one assembly while a different subset leads to the formation of the other (Figure 1). It suggests that these incubations are complex.

NaCl stress for Sec bodies and Osmotic stress for Stress granules: Second, Addition of NaCl to the growing medium triggers the formation of both stress assemblies. But Sec body formation requires the specific presence of NaCl. Only increasing NaCl led to a substantial formation of Sec bodies. Indeed, neither addition of KCl, Na-acetate nor osmotic shock (sucrose) induces their formation, suggesting that it is not downstream of osmotic stress. We have also shown that NaCl addition activates the SIKs.

Conversely, stress granule formation appears to be formed by the non-differential addition of salt (+150mM of NaCl, KCl, Na-acetate, LiCl, CaCl2), and addition of sucrose also leads to their formation, supporting that their formation is downstream of osmotic (and/or hypertonic) stress. However, we have shown that KRB and SCHNSO do not lead to a decrease in cell diameter (no apparent shrinkage). It suggests that osmotic salt stress activates specific pathways that still need to be fully elucidated. Osmotic stress induced stress granule formation falls in the category of non-canonical pathway as it appears to be independent of eIF2alpha phosphorylation (Aulas et al., 2017; Kedersha et al., 2002). The KRB and SCHNSO induced stress granule formation in S2 cells are also independent of the phosphorylation of this factor as its inhibition does not modulate the response. This suggests that this is also non canonical. In other systems, RocA and PatA have been involved and it could be interesting to test these enzymes in our system (Aulas et al., 2017; Kedersha et al., 2002).

Furthermore, the formation of stress granules upon KRB and SCHNSO appears to involve calmodulin, a protein that is the most elevated in KRB S2 cell stressed (Chapter 4).

Figure 1: The formation of Sec body and stress granules is governed by different intracellular pathways. Sec body formation is triggered by UPR activation combined to SIK activation, mimicked in the dish by addition of DTT and 100 mM NaCl. Stress granule formation is downstream of osmotic/ hypertonic stress.

This is relevant as calcium binding proteins are now clearly components of stress granules in mammalian cells (Markmiller et al., 2019; Marmor-Kollet et al., 2020). Their recruitment is in line with the finding that modulating the intracellular calcium level has an impact on stress granule formation. Taken together, the same cells stressed by the same extracellular conditions form two assemblies through different and non-overlapping intracellular pathways. This illustrates the complexities of the stress response in general and may explain why the two assemblies remain distinct instead of collapsing into a single structure.

Identification of the Sec body proteome
The fact that they remain distinct has allowed the identification of the Sec body proteome by the APEX strategy by tagging Sec24AB (a known component of the Sec bodies). We identified 52 proteins including large number of them being components of the ER, such as Sec16 (the key Sec body marker and driver), other COPII subunits, as well as components functioning at the Golgi apparatus. Interestingly, in line with phase separation, most of the 52 identified components contain intrinsically disordered regions (IDRs) that are proposed to contribute to coalescence (Babu et al., 2012; Guillen-Boixet et al., 2020). Importantly, these components are neither regulated transcriptionally nor translationally, suggesting a re-location to Sec bodies that likely act as storage during the period of stress.

Previous research in our group has shown that Sec bodies formation is related to the inhibition of protein secretion in the early secretory pathway and proteins recruit into Sec bodies protect them from degradation (Zacharogianni et al., 2014). Recent study in our group on the formation of Sec body in mammalian INS-1 cells suggests that Sec body formation is the cause of the inhibition of ER exit (van Leeuwen et al., 2022), not the consequence. Whether this is also the case in S2 cells remains to be explored in more detail. The incorporation of key components functioning in this pathway may store components and modulates the secretion and endocytosis capacity of the cells during stress.

More specifically, we found PARP16 and GM130 (that we validated) but also Hrs, Homer, Lasp, and Rox8.
-GM130 localized at ERES and cis-Golgi in Drosophila cells (Kondylis and Rabouille, 2009). In mammalian cells, GM130 forms a complex with p115 on the Golgi cisternae to help with building and maintaining Golgi stack, a similar complex (Kondylis and Rabouille, 2003; Nakamura et al., 1997). GM130 contain IDRs and this could contribute to its Sec body recruitment.
-Hrs is homologous to yeast Vps27p (vacuolar protein sorting), which regulates protein trafficking from a perivacuolar compartment to the vacuole in yeast (R C Piper et al., 1995). Drosophila and mammalian Hrs regulate inward budding of endosome membrane and (Multivesicular body) formation (Thomas, 2002). In both drosophila and mammalian cells, endosome enlarged phenotype were observed upon Hrs depletion or mutation (Thomas, 2002). Why Hrs is recruited into Sec bodies and whether Sec body formation correlates with inhibition (or modulation) of inward budding in late endosomes during stress still need further investigation.
-Homer colocalizes with an ER marker Bip suggested its location to the cytoplasmic face of the ER compartment (Thierry et al., 2002). Homer proteins are key components modulating calcium- and glutamate-induced calcium release from ER stores in neurons (Thomas, 2002; Xiao et al., 2000). Interestingly, as mentioned above, calcium signaling through calmodulin may be involved in stress granule formation in KRB and SCHNSO stress, suggesting that Homer recruitment to Sec bodies may participate to their formation through Ca2+ release from the ER. Depletion of this factor will help understanding its role in the formation of both assemblies.
-Drosophila Lasp is an actin-binding protein (Suyama et al., 2009) which highly similar to vertebrate and Caenorhabditis elegans Lasp (Chenard et al., 1998; Chew et al., 1998). Lasp consists of an N-terminal LIM domain, followed by two nebulin-like repeats, a spacer region and a C-terminal SH3 domain (Grunewald and Butt, 2008). SH3 domains (60 amino acids) are one of the most prevalent families of modular binding domains which present on signaling proteins. These 60 amino acid domains are evolutionarily conserved protein-protein interaction domains (Musacchio et al., 1992). SH3 domains mediate interactions that have been showed that are able to drive phase separation. For example, the linker between the first two SH3 domains in Nck enhances phase separation of Nck with N-WASP (Banjade et al., 2015). Whether SH3 domain of Lasp contributes Sec body formation under KRB incubation still need further investigation.
-Last, Rox8 (TIA1 in mammals) is an RNA binding protein and established stress granule marker in mammalian cells (Waris et al., 2014). It is now identified as a binding partner of Sec24AB specifically in KRB. Upon KRB, Rox8 localises in stress granules that are found in a close proximity to Sec bodies (marked by Sec16) near the ERES. This is in line with results previously described in (Zacharogianni et al., 2014) (Aguilera-Gomez et al., 2017) showing that these two assemblies form side by side but remain distinct one another.

Taken together, the identification of the Sec body proteome increases our knowledge of Sec bodies but also of other assemblies forming in their vicinity, such as stress granules (Rox8) but also Hrs containing coalescences that overlap with Sec16 without complete colocalization. This technique opens the possibility for identifying the Sec body component in high salt condition. It will be interesting to figure out whether any of these newly identified Sec body proteins are playing a role in controlling Sec body formation and other stress assemblies and how they modulate membrane trafficking steps.