Analysis of the Correlation between Stress Granule Assembly and Nucleus/Cytoplasm Localization of hnRNP A1, HuR and TIA1 During Arsenite-Induced Oxidative Stress

Eukaryotes tend to inhibit the global translation and modulate the synthesis, refolding, modification, turnover, kinetics of protein, in order to maintain the homeostasis of intracellular compartments and improve cell survival (Sfakianos et al., 2016; Mahboubi and Stochaj, 2017). The assembly of stress granules (SGs), a type of RNA- containing particle structure without membranous border in the cytoplasm, was considered as one conserved cell protective approach, targeting the transient harmful environmental assaults, such as oxidative stress, heat shock and virus infection (Thomas et al., 2011; Lloyd, 2016; Mahboubi and Stochaj, 2017). SGs harbor translationally arrested translation preinitiation complexes (PICs), and contain higher concentration of RNA transcripts and protein components, including eukaryotic translation initiation factor 2α (eIF2α), GAP SH3 domain-binding protein 1 (G3BP1), staphylococcal nuclease and Tudor domain containing 1 (SND1), heterogeneous nuclear ribonucleoprotein A1 (hnRNP A1), Hu Antigen R (HuR) and T cell intracellular antigen 1 (TIA1) (Thomas et al., 2011; Panas et al., 2016; Protter and Parker, 2016).

Several factors, including cell characteristics, stress type, time and intensity of stimulus, and post-translational modification of protein components, influence the accumulated size/number, and duration of SG structure (Kedersha et al., 2000; Buchan, 2014; Protter and Parker, 2016). Pathogenesis and therapeutic strategies of cancer and some other clinical diseases, such as ischemia, amyotrophic lateral sclerosis (ALS) and Alzheimer’s disease (AD), were also linked to aberrant SG biology (Buchan, 2014; Anderson, et al., 2015; Protter and Parker, 2016; Alberti et al., 2017; McCormick and Khaperskyy, 2017).

The assembly and disassembly of SGs in mammalian cells were reported to be related to several biological processes or cellular activities, such as autophagy, apoptosis, mitochondrial stress and antiviral immune response (Takahashi et al., 2013; Fu et al., 2016; Monahan et al., 2016; McCormick and Khaperskyy, 2017). It is meaningful to investigate the dynamic distribution of stress-associated protein in the SG aggregation and nucleus/cytoplasm localization during stress. Previously, we found that SND1-containing SGs actively communicate with the nuclear and cytosolic pool of HeLa cells (Gao et al., 2015). Here, in order to further investigate the correlation between highly dynamic SG formation and nucleo-cytoplasmic transport in HeLa cells exposed to arsenite-induced oxidative stress, we focused on three SG-associated protein components, namely hnRNP A1, HuR and TIA1, which are primarily nuclear.

Dulbecco’s minimal essential medium (DMEM, Invitrogen Life Technologies) with 10% fetal bovine serum (FBS) was used for the culture of HeLa cells.

Oxidative stress was induced to the treatment of 0.5 mM sodium arsenite. In addition, 10 μM Importazole (IPZ, Millipore/Merck, Germany) was utilized for the alteration of nucleus/cytoplasm location of targeting proteins

Two expression plasmids of RFP-hnRNPA1 and RFP-HuR were kindly provided by Prof. John Goodier (University of Pennsylvania School of Medicine, USA). Plasmid encoding GFP-TIA1 were kindly provided by Prof. Tom C. Hobman (McGill University, Montreal, Canada). Lipofectamine 2000 (Invitrogen, Barcelona, Spain) was used to transfect the above plasmids into HeLa cells, according to the manufacturer’s instructions.

4’,6-diamidino-2-phenylindole (DAPI, Sigma-Aldrich, USA) was used for staining the nucleus region of HeLa cells. The image data was collected via an inverted research microscope Leica. The ratio of cells with the nucleus-localized signal, cytoplasm-localized signal, or full cell-localized signal of hnRNP A1, HuR, and TIA1, per all cells was calculated, respectivly.

A total of 50~100 HeLa cells were randomly scored in each experiment. Calculation of the percentage of targeting granule positive cells was performed. SG number in selected cells was analyzed, and cell ratio of four different granule number ranges, namely 1~10, 11~20, 21~30 and >30 per cell, were calculated, respetively. Independent-Sample Student’s T Test was performed via SPSS 16.0 software. P value less than 0.05 means the existence of statistical significant difference.

Cytoplasmic SG formation was linked to nucleus/cytoplasm localization of hnRNP A1, HuR and TIA1 during arsenite induced-oxidative stress

In order to study the relationship between the nucleus/cytoplasm localization and cytoplasmic SG formation, three nucleocytoplasmic shuttling protein components, including hnRNP A1, HuR and TIA1, were targeted. HeLa cells were cultured and transiently transfected with plasmids of RFP-hnRNP A1, RFP-HuR, GFP-TIA1, rspectively. The RFP and GFP vectors were included as control. Treatment of 0.5 mM sodium arsenite for 1 hour was used to induce the oxidative stress of HeLa cell. As shown in Figure 1A, RFP-tagged hnRNP A1 protein was mainly located in the nucleus section of cells under normal condition. Upon the oxidative stress, we observed the formation of hnRNP A1 potitive (hnRNP A1+) SGs in the majority of HeLa cells (Fig. 1A, Fig. 2A, P<0.05) and the alteration of nucleus/cytoplasm localization signal (Fig. 2B). Increased hnRNP A1 fluorescence signal within cytoplasm was detected from 3% of normal cells to the 28% of stressed cells; in contrast, 87% of cells with strong hnRNP A1 signal within nucleus reduced to 50% during stress (Fig. 2B).

Furthermore, we also found that HuR-containing SGs were aggregated in the cytoplasm of HeLa cells exposed to oxidative stress (Fig. 1B; Fig. 2C, P<0.05), as expected. Similarly, 11% with strong cytoplasmic HuR signal under normal condition increased to 18% under stress condition, whereas 56% with strong nuclear HuR signal reduced to 49% upon the oxidative stress (Fig. 2D).

The same trend was also observed for TIA1 protein (Fig. 1D; Fig. 2E, P<0.05; Fig. 2F). However, this phenomenon was not detectable for the vector control of GFP (Fig. 1C) and RFP (Fig. 1E). These suggested that the cytoplasmic aggragation of hnRNP A1, HuR and TIA1-containing SGs was linked to the nucleus/cytoplasm localization during arsenite induced-oxidative stress.

Next, we aim at analyzing whether altered nucleus/cytoplasm localization can influence the SG formation during stress. HeLa cells were cultured in the presence (+) or absence (-) of 10 μM Importazole (IPZ), a small molecule inhibitor of the transport receptor importin-β (Soderholm et al., 2011). As shown in Figure 3A and Figure 4A, only a slight trend for the alteration of nucleus/cytoplasm localization was observed for hnRNP A1, after the treatment of IPZ. 50% with strong nuclear hnRNPA1 signal under normal condition reduced to 41%; whereas 28% with strong cytoplasmic hnRNPA1 signal increased to 34% (Fig. 4A). Moreover, there is no difference for SG size (Fig. 3A) and the portion of cells with hnRNP A1 positive (hnRNP A1+) SGs (Fig. 4B) in HeLa cells treated with IPZ or not. We also failed to observe the statistical significant difference for the cells with different granule numbers, including 1~10, 11~20, 21~30, >31, between the IPZ (-) and IPZ (+) group (Fig. 4C). These suggested that IPZ-mediated slight alteration of nucleus/cytoplasm

localization does not significantly affect the formation of hnRNP A1-containing SGs in HeLa cells during the arsenite-induced oxidative stress.

With regards to HuR protein, we are surprised to observe an increased nuclear aggregation in stressed HeLa cells (Fig. 3B and Fig. 4D). Even though no difference for the HuR positive (HuR+) SG formation (Fig. 4E) and HuR+ SG size (Fig. 4B), decreased portions of HeLa cells with the granule number of 11~20, 21~30, was observed in the group of IPZ (+), compared with IPZ (-) group (Fig. 4F, P<0.05). In addition, there are similar change for the nucleus/cytoplasm localization (Fig. 4G), the portions of HeLa cells with TIA1 positive (TIA1+) SGs (Fig. 4H), TIA1+ SG size (Fig. 3D) and different granule numbers (Fig. 4I). No change was observed in the control group of RFP (Fig. 4C) and GFP (Fig. 4E). Overall, these data suggested that transport receptor importin-β pathway seems to be involved in the nuclear import of hnRNP A1, rather than HuR and TIA1. The slightly enhanced cytoplasmic accumulation of hnRNP A1 can not influence the formation of hnRNP A1-containing SGs during oxidative stress.

Nucleocytoplasmic shuttling system, including nuclear localization signal (NLS), nuclear export signal (NES) and importin-α/ importin-β pathway, is essential for the nuclear transport of many proteins among eukaryotes (Cardarelli, 2017). The environmental stress was linked to the dynamic flux of proteins between the nuclear and cytoplasmic compartment. For instance, some stress conditions, such as starvation and heat shock, inhibit the importin α/ importin-β-mediated nuclear import of the small GTPase Gsp1p (Stochaj et al., 2000). In addition, nuclear transport factors, such as Importin-α1, Importin-α4 and Importin-α5, were reported to be recruited to SG structure (Mahboubi et al., 2013). Meanwhile, a number of nuclear protein components are involved in the formation of cytoplasmic SGs (Thomas et al., 2011; Protter and Parker, 2016). In the present study, we quantitatively analyzed the correlation between SG assembly and nucleus/cytoplasm localization of nuclear hnRNP A1, HuR and TIA1 proteins during arsenite-induced oxidative stress.

hnRNP A1, a rapid nucleo-cytoplasmic shuttling protein and alternative splicing factor, binds nascent RNA polymerase II manuscripts in cytoplasm (Pinol-Roma and Dreyfuss, 1992), and functions in the pre-mRNA processing, RNA metabolism, cellular apoptosis, stress adaption, post-transcription modulation, microRNAs processing and telomere maintenance (Jean-Philippe et al., 2013). hnRNP A1 was found to be localized into SG structure during stress, and its recruitment into SGs depends on the existance of methylarginine residues within the arginine-glycine-glycine (RGG)-motif region (Guil et al., 2006; Wall and Lewis, 2017). In addition, RanGTP-binding nuclear transport receptor transportin1 was invloved in the nuclear reimport of hnRNP A1 (Rebane et al., 2004). Guil et al. (2006) reported that Mnk1/2-mediated hnRNP A1 phosphorylation during stress fails to bind sufficiently transportin 1, which is helpful for the cytoplasmic aggregation of hnRNP A1 into SGs. Here, we found that the arsenite-induced oxidative stress induces the increased cytoplasmic signal, but the decreased nuclear signal of hnRNP A1. The inhibition of nuclear import or enhanced nuclear export may contribute to the cytoplasmic accumulation of hnRNP A1, when HeLa cells undergo the stress of sodium arsenite. Similarly, the treatment of osmotic shock can induce the cytoplasmic accumulation of hnRNP A1 in NIH 3T3 cells through the serine-specific phosphorylation of a C-terminal F-peptide, and the activation of the mitogen-activated protein kinase kinase 3/6-p38 signaling pathway (Allemand et al., 2005). Previous FRAP results showed that hnRNP A1 protein moves continuously in and out of SG structure (Guil et al., 2006). The high kinetic behaviour allows the cells to respond quickly to the adverse stress, by way of dynamically influencing and equilibrating the subcellular distribution of hnRNP A1.

We also utilized a small molecule inhibitor of the transport receptor importin β (IPZ) to analyze the effect of altered nucleus/cytoplasm localization on the assembly of hnRNP A1 granules. We observed the slightly increased cytoplasmic hnRNP A1 signal after the treatment of10 μM IPZ, under the normal (data not shown) and stress conditions. However, we failed to observe the change of hnRNP A1 granule formation. It is pissible that the increased amount of cytoplasmic hnRNP A1 protein was not up to the level of affecting the SG assembly. Previously, we found that the inhibition of nuclear export of SND1 protein could influence the efficient aggreation of SND1 granule in the cytoplasm of HeLa cells (Gao et al., 2015). The effect of stress-responsive nucleus/cytoplasm localization on the SG formation may be protein-specific.

TIA-1 recognizes adenine/uridine rich elements (AREs) within 3’-untranslated regions (3’-UTR) of RNA, and the prion-like aggregation of TIA1 was required for the formation of mammalian SGs under stress conditions (Kedersha et al., 2000; Gilks et al., 2004; Waris et al., 2014). TIA1 exhibited the similiar cellular kinetics with hnRNP A1 (Guil et al., 2006). Here, we observed the oxidative stress-induced slightly decreased nuclear accumulation of TIA1, however, IPZ treatment fails to influence the nucleus/cytoplasm localization of TIA1. HuR, a classical member of embryonic lethal abnormal vision (ELAV) family, recognizes AU-rich elements (AREs) sequences of targeting mRNAs, and is involved in the regulation of RNA metabolism (Grammatikakis et al., 2017). HuR shuttles between nucleus and cytoplasm, and heat shock inhibits the interaction between HuR and cytoplasmic mRNA (Gallouzi et al., 2000). HuR locolized into SGs as well (Kedersha and Anderson, 2002), and RanGTP-binding nuclear transport receptor Transportin 2 was shown as nuclear import factor of HuR (Guttinger et al., 2004). Even though a slightly increased cytoplasmic signal of HuR was oberved under the oxidative stress, IPZ treatment did not alter the nuclear import of HuR but decrease the portions of HeLa cells with the granule number of 11~20, 21~30, which involved some unknown reasons.

Taken together, our study focuses on the association between stress granule formation and nucleus/cytoplasm location of hnRNP A1, HuR and TIA1. Cytoplasmic SG aggregation behavior is linked to the alteration of nucleus/cytoplasm localization of hnRNP A1. In response to unfavorable environmental stress, timely and effective dynamic distribution of specific stress-associated proteins in the section of nucleus, cytoplasm, and SGs is more likely to contribute to the minimization of the detrimental condition-induced cellular damage.

This work was supported by grants from the National Science Foundation for Distinguished Young Scholars of China (31125012), National Natural Science Foundation of China (31571380, 31701182, and 31670759), Innovation Team Development Plan of the Ministry of Education (IRT13085), Tianjin Enterprise Science and Technology Commissioner Project (18JCTPJC59400).

The authors declare they have no conflicts of interest.