Upon exposure to various environmental stresses such as arsenite, hypoxia, and heat shock, cells inhibit their translation and apoptosis and then repair stress-induced alterations, such as DNA damage and the accumulation of misfolded proteins. These types of stresses induce the formation of cytoplasmic RNA granules called stress granules (SGs). SGs are storage sites for the many mRNAs released from disassembled polysomes under these stress conditions and are essential for the selective translation of stress-inducible genes. Ras-GTPase-activating protein SH3 domain-binding protein 1 (G3BP1) is a component of SGs that initiates the assembly of SGs by forming a multimer. In this study, we examined the role of G3BP2, a close relative of G3BP1, in SG formation. Although single knockdown of either G3BP1 or G3BP2 in 293T cells partially reduced the number of SG-positive cells induced by arsenite, the knockdowns of both genes significantly reduced the number. G3BP2 formed a homo-multimer and a hetero-multimer with G3BP1. Moreover, like G3BP1, the overexpression of G3BP2 induced SGs even without stress stimuli. Collectively, these results suggest that both G3BP1 and G3BP2 play a role in the formation of SGs in various human cells and thereby recovery from these cellular stresses.
Upon exposure to various types of stress, such as arsenite, heat shock, hypoxia, and viral infections, mammalian cells activate protective mechanisms to evade the accumulation of DNA and protein alterations. Such protective mechanisms include the formation of stress granules (SGs), which are stress-inducible cytoplasmic granules containing various mRNAs (White et al. 2007; Anderson & Kedersha 2009; Buchan & Parker 2009; Holley et al. 2011). The formation of SGs is essential for the recovery from stresses and the inability to form SGs induces apoptosis (Kim et al. 2005; Arimoto et al. 2008).
The formation of SGs is associated with the disruption of polysomes, followed by the transfer of polysome-released mRNAs into SGs in an inactive state (Buchan & Parker 2009). Therefore, one of the main functions of SGs is transient storage of mRNAs during stress. Such storage of mRNAs by SGs is selective, and housekeeping genes are stored in SGs, whereas stress-responsive genes, such as genes encoding heat shock proteins, are able to escape from the storage and are selectively translated to repair stress-induced alterations (Kedersha & Anderson 2002; Stohr et al. 2006; Anderson & Kedersha 2009). During the recovery phase, SGs are rapidly dispersed, and the mRNAs stored in SGs have two fates; mRNAs can move back to polysomes to restart translation, or they can move to processing bodies, another type of RNA granule, in which they are degraded (Anderson & Kedersha 2008; Buchan & Parker 2009).
Stress granules contain several RNA-binding proteins, including Ras-GTPase-activating protein SH3 domain-binding protein 1 (G3BP1), TIA-1, eukaryotic initiation factor (eIF)3, eIF4E, eIF4G, poly-A binding protein (PABP), and several ribosomal proteins (rps6, rps3 etc.) (Kedersha et al. 1999, 2000; Kedersha & Anderson 2002; Kimball et al. 2003; Tourriere et al. 2003; Anderson & Kedersha 2006). These RNA-binding proteins are thought to play roles in SG-associated functions. Of them, G3BP1 plays a role in the formation of SGs, as evidenced by the fact that knockdown of G3BP1 reduces the assembly of SGs (Ghisolfi et al. 2012). Treatment of cells with arsenite induces the dephosphorylation of G3BP1 at Ser-149, and this dephosphorylation stimulates the multimerization of G3BP1 (Tourriere et al. 2003). Therefore, stress-induced multimerization of G3BP1 is likely to initiate SG formation.
The G3BP1 protein can be divided into five regions, the NTF2-like domain, the acidic region, the PXXP motif, and two RNA-binding motifs [an RNA recognition motif (RRM) and an arginine-glycine rich box (RGG)]. The NTF2-like domain and at least one RNA-binding motif are required for SG formation (Tourriere et al. 2003). In addition, it has been shown that G3BP1 has an endoribonuclease activity for certain RNAs, although it is unclear whether this activity is associated with functions associated with SGs (Gallouzi et al. 1998; Tourriere et al. 2001).
G3BP2 is a close relative of G3BP1, with a similar domain architecture, including the NTF2-like domain and two RNA-binding motifs (Kennedy et al. 2001; Irvine et al. 2004). Like G3BP1, G3BP2 is ubiquitously expressed and is recruited into SGs (French et al. 2002; Kobayashi et al. 2012). In this study, we examined the role of G3BP2 in the formation of SGs.
G3BP2b is recruited into SGs in various cells
To examine whether G3BP2 is also recruited into SGs, we transfected the FLAG-G3BP2b expression plasmid into HeLa cells, and the cells were then treated with a pro-oxidant, arsenite (Fig. 1). An immunofluorescent analysis showed that FLAG-G3BP2b was mainly localized in the cytoplasm of HeLa cells. Arsenite treatment induced the formation of SGs, as detected by PABP staining in cells, and FLAG-G3BP2b was recruited into these SGs. These results are consistent with the previous study showing that endogenous G3BP2 is recruited into SGs after treatment with arsenite or heat shock (Kobayashi et al. 2012).
Both G3BP1 and G3BP2 interact with USP10 and PABP
G3BP1 interacts with ubiquitin specific protease 10 (USP10) and PABP (Soncini et al. 2001; Sowa et al. 2009), and the interaction of G3BP1 with USP10 is important for the SG-mediated inhibition of apoptosis (manuscript submitted for publication). We therefore examined whether G3BP2 also interacts with USP10 and/or PABP. Cell lysates prepared from 293T cells were immunoprecipitated with anti-G3BP2 antibodies, and the immunoprecipitates were characterized by either anti-USP10 or anti-PABP. Either endogenous G3BP2 or G3BP1 interacted with both endogenous USP10 and PABP (Fig. 2A,B). G3BP1 and G3BP2 have RNA interaction motifs (RRM, RGG in Fig. 3). Therefore, we next examined whether the observed interactions are mediated by RNA bound to G3BPs. The 293T cells were transfected with FLAG-G3BP2b or FLAG-G3BP1 plasmids, and the cell lysates were treated with RNase before undergoing immunoprecipitation. FLAG-G3BP2b and FLAG-G3BP1 interacted with endogenous USP10 and PABP; however, the interactions of G3BPs with PABP but not with USP10 were abrogated by RNase treatment (Fig. 2C). These results indicate that G3BP1 and G3BP2 interact with PABP indirectly through bound RNAs, whereas they interact with USP10 without RNAs.
Next, we examined which domains of G3BP1 and G3BP2b were required for the interaction with USP10 or PABP (Fig. 3). The studies using deletion mutants indicated that the NTF2-like domain of both G3BP1 and G3BP2b is sufficient for their interaction with USP10 (Fig. 3A,C). A further deletion analysis of G3BP1 indicated that the N-terminal 46 amino acids of G3BP1 in the NTF2-like domain are required for the interaction with USP10 (Fig. 3B,D). However, the deletion mutant studies indicated that two regions of G3BP1 and G3BP2b (the NTF2-like domain and the PXXP motif) were independently sufficient for the interaction with PABP.
G3BP1 and G3BP2 double-knockdown impairs SG formation
To elucidate the role of G3BP1 and G3BP2 in SG formation, the amounts of G3BP2 and G3BP1 in cells were transiently reduced by small interfering RNAs (siRNAs) specific for human G3BP1 or G3BP2 RNA (G3BP1-siRNA-1, G3BP2-siRNA-1). A Western blot analysis showed that both single and double knockdown significantly reduced the expression of the G3BP2 and/or G3BP1 protein in 293T and HeLa cells (Fig. 4A). To measure the SG formation in these cells, they were treated with 0.5 mm sodium arsenite. Although single knockdown of either G3BP1 or G3BP2 partially reduced the number of SG-positive 293T cells, the double knockdown much more significantly reduced the number (Fig. 4B,C). In addition, whereas single knockdown of G3BP2 little affected SG formation in HeLa cells, the knockdown prominently reduced SG formation in G3BP1-knockdown cells. These results indicate that either G3BP1 or G3BP2 is required for arsenite-induced SG formation in 293T and HeLa cells.
To confirm the specificities of the siRNAs used above, we designed a rescue experiment. We first established stable G3BP2-knockdown 293T cells by lentivirus-mediated transduction of G3BP2 short-hairpin RNA (shRNA) and selection with puromycin. The puromycin-resistant cells were then transiently transfected with G3BP1-siRNA-2, which targets the 3′-untranslated region (UTR) of G3BP1 RNA. A Western blot analysis confirmed the reduced expression of both G3BP1 and G3BP2 in the double-knockdown cells (Fig. 5A). These double-knockdown cells were then transfected with a FLAG-G3BP1 plasmid, the transcript of which does not have the UTR targeted by the G3BP1-siRNA-2. A Western blot analysis confirmed the expression of wild-type G3BP1 in the FLAG-G3BP1-transfected double-knockdown cells (Fig. 5A). Then, these cells were examined for SG formation (Fig. 5B,C). Exogenous G3BP1 in the double-knockdown cells efficiently rescued the SG formation to a level equivalent to that of wild-type 293T cells. In contrast, the SG formation was not rescued in cells transfected with one of three G3BP1 mutants with a deletion in the NTF2-like domain.
The double-knockdown cells were also transfected with a shRNA-resistant G3BP2b plasmid (Fig. 6). Wild-type G3BP2 protein rescued the SG formation, and the resulting level was equivalent to that in wild-type 293T cells. However, transfection with either of two G3BP2 mutants with a deletion of the NTF2-like domain could not rescue the SG-forming activity (Fig. 6). In addition, neither G3BP1(1–138) nor G3BP2b(1–138) rescued the SG formation, thus indicating that the NTF-like region of G3BP1 or G3BP2 is not sufficient for SG formation (Figs 6, 7). These results indicate that the NTF2-like domain and the rest of the C-terminal region of G3BP2 containing RNA interaction motifs are both required for SG formation. We also noticed that the expression of G3BP2b(139–449), but not G3BP2b(296–449), reduced the number of SG-positive double-knockdown cells (Fig. 6C). This might indicate that the acidic region of G3BP2b has an inhibitory effect on SG formation because there appeared to be an inhibitory effect on the remaining endogenous G3BPs.
Overexpression of G3BP2 without stress induces SG formation
A previous study showed that the overexpression of G3BP1 in cells without stress stimuli induced SG formation (Tourriere et al. 2003). We therefore examined whether G3BP2 also has a similar activity. The overexpression of G3BP2b in 293T cells also induced SGs, but the number of SG-positive cells (14%) was less than that (27%) of cells overexpressing G3BP1 (Fig. 8). In addition, the size of the SGs induced by G3BP2b was smaller than those induced by G3BP1. These results indicate that G3BP2 by itself has SG-forming activity even in the absence of stress, but its activity is lower than that of G3BP1.
A previous study suggested that the multimerization of G3BP1 after dephosphorylation of Ser-149 is required for SG formation and that the activity is dependent on the NTF2-like domain (Tourriere et al. 2003). G3BP2 may also form a multimer, because it also has the Ser-149 adjacent to the NTF2-like domain. To examine this possibility, the 293T cells were transfected with Myc-G3BP2b together with either FLAG-G3BP2b or FLAG-G3BP1, and the cells were treated with arsenite. The cell lysates were then immunoprecipitated with anti-Myc, and the immunoprecipitates were characterized by anti-FLAG antibodies. Myc-G3BP2b immunoprecipitated FLAG-G3BP2b as well as FLAG-G3BP1 before arsenite treatment and the levels were equivalent to those observed after arsenite treatment (Fig. 9A,B). These results suggest that G3BP2 also forms homo- and hetero-multimers with G3BP1 that are likely to be required for SG formation.
G3BP1 and G3BP2 have 59% amino acid similarity, and share similar domain structures, including the NTF2-like domain and two RNA-binding motifs (RRM, RGG) (Fig. 3, Kennedy et al. 2001; French et al. 2002; Irvine et al. 2004). We therefore speculated that G3BP2 also plays a role in SG formation. In this study, we showed that G3BP2 is recruited into SGs in 293T and HeLa cells and that either G3BP2 or G3BP1 is required for SG formation induced by arsenite. It should be noted that some tissues studied in mice, such as the brain, muscle and small intestine, dominantly express the G3BP2 protein, whereas G3BP1 was undetectable (Kennedy et al. 2001). As a result, in these tissues, G3BP2 might be the main player in SG formation and its associated functions.
G3BP2 formed a homo- and a hetero-multimer with G3BP1 (Fig. 9). Tourriere et al. (2003) presented evidence that the augmented multimerization of G3BP1 induced by stress-induced dephosphorylation at Ser-149 initiates SG formation. Collectively, the present study suggests that G3BP1 and G3BP2 form homo- and hetero-multimers to induce SGs. It should be noted that G3BP2 also has the Ser-149 adjacent to the NTF2-like domain and that the amino acid sequences surrounding the Ser-149 of G3BP1 and G3BP2 are highly conserved. Therefore, similar to G3BP1, the dephosphorylations of G3BP2 at Ser-149 induced by stressors might also augment its homo- and hetero-multimerization with G3BP1 to initiate SG formation.
The overexpressions of G3BP2b without stress-induced SG formation in the 293T cells; however, the formation was approximately half of that induced by G3BP1 (Fig. 8). Although the mechanism underlining the difference between G3BP1 and G3BP2 is unclear at the present moment, the recruitment of RNAs by G3BP2 overexpression might be less than that of G3BP1. Therefore, further analysis is required to determine what causes this difference between G3BP1 and G3BP2 and whether or not the difference has any physiological significance.
The overexpression of the acidic region of G3BP2b without the NTF-like domain reduced SG formation (Fig. 6). A previous study showed that the acidic region of G3BP1 showed inhibitory activity for SG formation (Tourriere et al. 2003). Therefore, the acidic region of G3BP1 and G3BP2 might interact with a factor promoting SG formation.
Both G3BP1 and G3BP2 interact with USP10 through the NTF2-like domain. We recently observed that knockout of USP10 in murine embryonic fibroblasts reduces SG formation and augments apoptosis in cells treated with arsenite (submitted for publication). As a result, through its interaction with USP10, G3BP2 might play a role in inhibiting apoptosis under stress conditions. Further studies are required to clearly elucidate the functions of G3BP2 under stress conditions.
Cells and culture conditions
The 293T and HeLa cells were originated from human embryonic kidney and human cervical cancer, respectively, and the cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS), 4 mM l-glutamine, penicillin (50 μg/mL), and streptomycin (50 μg/mL) at 37 °C in 5% CO2.
pFLAG-CMV2 is an expression vector encoding a protein with an in-frame FLAG epitope tag at its N-terminus (Sigma-Aldrich). The cDNA fragments of G3BP2b and their mutants were amplified from the pcDNA3.1(+)-G3BP2b plasmid by polymerase chain reaction (PCR), and the fragments were inserted into the cDNA cloning vector pCR-BluntII-TOPO (Invitrogen) and were then cloned into the BglII and SalI sites of pFLAG-CMV2. pcDNA6-mycHisA is an expression vector encoding a protein with an in-frame Myc-epitope and 6-histidine-epitope tag at its C-terminus (Invitrogen). The G3BP2b cDNA fragment with a deletion of the stop codon was cloned into the KpnI and XhoI sites of pcDNA6-mycHisA, and the construct was designated as Myc-G3BP2b. pLKO.1-puro-shG3BP2 expresses a shRNA targeting human G3BP2 RNA (MISSION shRNA; Sigma-Aldrich). This G3BP2 shRNA targets both G3BP2a and G3BP2b RNA.
Small interfering RNAs
The siRNAs specific for human G3BP1 RNA (Stealth Selected RNAi™, siRNA ID: HSS115446) and human G3BP2 RNA (ID: HSS114988) and negative control siRNAs were purchased from Invitrogen. G3BP1 siRNA-2 targeting the 3′-untranslated region of human G3BP1 RNA (ID: SASI_Hs01_00045804) and universal negative control siRNA were purchased from Sigma-Aldrich. Transfection was carried out with 100 pmol siRNAs using Lipofectamine 2000 or Lipofectamine RNAiMAX reagents according to the manufacturer's protocol (Invitrogen).
Establishment of G3BP2-knockdown cells
Recombinant lentiviruses were generated by transfecting pCAG-HIVgp, pCMV-VSV-G-RSV-Rev, and pLKO.1-puro-shG3BP2 into 293T cells using FuGENE 6 (Roche). Forty-eight hours after the transfection, the culture supernatant was collected and used to infect 293T cells (2 × 105) in a final volume of 2 mL DMEM/10% FBS containing 8 μg/mL polybrene. At 24 h after the infection, the cells were cultured in the presence of 2 μg/mL of puromycin for 3 days. The expression of the G3BP2 protein in the selected cells was measured by a Western blot analysis.
The nucleotide sequences of G3BP2 cDNA targeted by shRNA were mutated by PCR-based site-directed mutagenesis. pFLAG-CMV2-G3BP2b was used as a template for PCR amplification. The nucleotide sequences of the primers were as follows; 5′-CCACAAGGTGTTGTCTCTCAACTTCAGTGAATGTCAT-3′ and 5′-TTCAGAGACAACACCTTGTGGTGTATATCATTTTGGCC -3′ (the mutated nucleotides were underlined). The mutant G3BP2b cDNA fragment was inserted into the pCR-BluntII-TOPO vector, and then the fragment was inserted into the BglII and EcoNI site of the pFLAG-CMV2 expression vector.
Western blot analysis
Cells were treated with sodium dodecylsulfate (SDS) buffer [2% SDS, 62.5 mm Tris-HCl (pH 6.8), 20% glycerol, 0.01% bromophenol blue, 50 mm dithiothreitol], sonicated, and heated at 95 °C for 5 min. After centrifugation, the supernatants (20 μg) were size-separated on 6–12% polyacrylamide gels containing SDS, and they were then electronically transferred to PVDF membranes. The membranes were incubated with the primary antibody, followed by the corresponding secondary antibody labeled with peroxidase, and the antibody binding was visualized using an ECL Western blotting detection system (GE Healthcare). The primary antibodies used were anti-G3BP1 (BD Bioscience Pharmingen), anti-FLAG (M2 Monoclonal Antibody; Sigma), anti-G3BP2 (Bethyl Laboratories Inc.), anti-USP10 (Bethyl Laboratories Inc.), anti-c-Myc (Sigma), anti-β-actin (Santa Cruz), and mouse anti-α-tubulin monoclonal antibodies (Calbiochem).
Immunostaining analysis and measurement of SG-positive cells
To determine the subcellular localization of PABP, G3BP1, G3BP2, and their mutants, 293T cells were cultured on glass slides on a six-well culture plate for 24 h and then the cells were transfected with G3BP-siRNA (20 nm) or control siRNA using the lipofection method (Lipofectamine 2000). At 24 h after the transfection, the cells were transfected with either the pFLAG-G3BP1 or its deletion mutants using the lipofection method. Twenty-four hours after this transfection, the cells were treated with 0.5 mm sodium arsenite for 60 min. The cells were then washed with PBS and fixed with 4% formaldehyde for 15 min at room temperature. After washing with PBS, the cells were permeabilized with 1% Triton X-100 for 5 min at room temperature, washed with PBS, and dried for 20 min at room temperature. The fixed cells were then incubated with 3% bovine serum albumin (blocking solution) for 60 min at room temperature and then the cells were incubated with the primary antibody for 60 min at room temperature. After being washed with 0.2% Triton X-100 three times, the cells were incubated with Alexa594-labeled anti-mouse IgG (Molecular Probes, Eugene, OR, USA), Alexa488-labeled anti-rabbit IgG (Molecular Probes) and Hoechst33258 for 60 min at room temperature. The stained cells were examined by fluorescent microscopy (BZ-8000; KEYENCE). SG formation was examined by anti-PABP staining. After examining more than 150 cells, cells with SGs more than 2 μm in size was judged as SG-positive cells, and the proportion of cells with SGs (%) was calculated. The primary antibodies used were anti-FLAG (M2 monoclonal antibody; Sigma), rabbit anti-FLAG (Sigma), mouse anti-G3BP1 (BD Bioscience Pharmingen), mouse anti-PABP (Sigma), and rabbit anti-PABP (Abcam).
The pFLAG-CMV2-G3BP1, pFLAG-CMV2-G3BP2b plasmids or the deletion mutants (1 μg) were transiently transfected into 293T cells by the lipofection method (Fugene6). At 48 h after the transfection, the cells were treated with lysis buffer (1% Nonidet P-40, 125 mm Tris-HCl, pH 7.2, 150 mm NaCl, 1 mm EDTA, pH 8.0, 1 mm phenylmethylsulfonyl fluoride, 0.2 μg/mL aprotinin) and cell lysates were cleared by centrifugation, and immunoprecipitated with the anti-FLAG antibody. Anti-normal mouse IgG (Santa Cruz) and anti-normal rabbit IgG (Santa Cruz) were used as control antibodies for immunoprecipitation. The immunoprecipitated proteins were characterized by a Western blot analysis using the corresponding primary and secondary antibodies.
We would like to thank Dr Masaya Fukushi for providing the pcDNA3.1(+)-G3BP2b, pcDNA3.1-mycHisB and pcDNA6-mycHisA-G3BP2b plasmids. We also wish to thank Dr Hiroyuki Miyoshi at RIKEN Tsukuba Institute for providing the lentivirus plasmids. Finally, we express our gratitude to Misako Tobimatsu for her excellent technical assistance.