Effect of the glycine-rich domain in GAREM2 on its unique subcellular localization upon EGF stimulation

In mammals, there are two subtypes of Grb2-associated regulator of Erk/MAPK (GAREM), an adaptor protein that functions downstream of the cell growth factor receptor. GAREM1 is ubiquitously expressed, whereas GAREM2 is mainly expressed in the brain. However, the precise mechanism of the translocation of each GAREM subtype in growth factor-stimulated cells is still unclear. In this study, immunofluorescence staining with specific antibodies against each GAREM subtype and time-lapse analysis using GFP fusion proteins were used to analyze the subcellular localization of each GAREM subtype in a cell growth stimulus-dependent manner. We also biochemically analyzed the correlation between its subcellular localization and tyrosine phosphorylation of GAREM2. We found that endogenously and exogenously expressed GAREM2 specifically aggregated and formed granules in NGF-stimulated PC-12 cells and in EGF-stimulated COS-7 cells. Based on the observed subcellular localizations of chimeric GAREM1 and GAREM2 proteins, a glycine-rich region, which is present only in GAREM2, is required for the observed granule formation. This region also regulates the degree of EGF-stimulation-dependent tyrosine phosphorylation of GAREM2. Our results, showing that aggregation of GAREM2 in response to EGF stimulation is dependent on a glycine-rich region, suggest that GAREM2 aggregation may be involved in neurodegenerative diseases.

Page 2 of 11 Nishino et al. Cell Mol Biol Lett (2021) 26:16 Grb2-associated regulator of Erk/MAPK (GAREM) is an adaptor protein that functions downstream of the EGF receptor signaling pathway. It is tyrosine phosphorylated upon EGF stimulation and associates with Grb2 via a proline-rich region in the molecule [8]. There are two variants of GAREM, which have different expression patterns in mammals; GAREM1 is ubiquitously expressed, whereas GAREM2 expression is brainspecific [9].
Analysis of GAREM2 knockout mice showed that while the loss of GAREM2 does not cause significant changes in brain function itself, these knockout mice exhibited emotional alterations, such as less anxiety, mildly increased social approaching behavior, and longer latency to immobility as compared to wild-type mice (WT) [10]. Furthermore, neurite outgrowth in primary cultured neurons was suppressed in neurons derived from GAREM2 knockout mice. In addition, neurite outgrowth of SH-SY5Y cells after IGF-I treatment was inhibited by knockdown of GAREM2 expression [9].
In this study, we focused on GAREM2-specific aggregated granule formation in EGFstimulated cells, and we found a region of GAREM2 that is required for both its subcellular localization and tyrosine phosphorylation.

Materials and methods
Cell culture, transfection and reagents COS-7 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS), 100 μg/ml of streptomycin, and 100 units/ml of penicillin. PC-12 cells were cultured in Dulbecco's modified minimal essential medium containing 10% calf serum and 5% horse serum. COS-7 cells were transfected with each vector by electroporation using a Gene-Pulser (Bio-Rad). The cells were serum-starved for 16 h, and 100 ng/ml of EGF (Sigma) or 50 ng/ml of NGF (Upstate) dissolved in a serum-free medium was added.

Expression plasmids
FLAG-tagged expression vectors of the GAREM family and its derivatives were constructed as previously described [8,9]. To express amino-terminal GFP-fused GAREM1 and GAREM2 in COS-7 cells, the full length of each GAREM cDNA was cloned into a pEGFP-C3 vector (Takara Clontech) in frame. To generate chimeric proteins, we first introduced the appropriate restriction enzyme sites into the cDNA of each GAREM subtype by site-specific mutagenesis by PCR. Next, the five chimeric protein expression vectors shown in Fig. 2 were constructed to ligate the respective DNA fragments. The DNA sequences were confirmed using a CEQ8000 DNA Sequencer (BECKMAN).

Immunoprecipitation and immunoblot analysis
The following experiments were performed at 0-4 °C. The transfected cells were lysed in lysis buffer containing 20 mM Tris-HCl (pH 7.5), 1 mM EDTA, 10 mM dithiothreitol, 1% Triton X-100, 150 mM NaCl, 10 mM NaF, 1 mM Na 3 VO 4 , and a complete protease inhibitor mixture (Roche Applied Science). The lysate was used as total cell lysate (TCL). For immunoprecipitation experiments, the total cell lysate was centrifuged, and the supernatant was incubated for 2 h with either the primary antibody or an anti-FLAG affinity gel (Sigma). The beads were then washed three times with a lysis buffer. The processed samples were separated by SDS-PAGE and analyzed by immunoblotting as described previously [8].

Fluorescence microscopy analysis and confocal microscopy
Transfected and nontransfected cells were fixed with 5% formaldehyde in PBS for 10 min, washed with PBS, and permeabilized with 0.1% Triton X-100 in PBS (PBST) for 10 min. Following a blocking step with 3% bovine serum albumin in PBST for 30 min, the primary antibodies (anti-GAREM, anti-FLAG or anti-EGFR) were applied for 1 h. After washing with PBST, the cells were treated with the appropriate secondary antibodies conjugated with Alexa fluorescent dyes (Molecular Probes) for 45 min. The nuclei were simultaneously stained with 2 μM Hoechst 33,342 (Molecular Probes). Then, the cells were rinsed with PBST and mounted onto microscope slides with ProLong Antifade reagents (Molecular Probes). Fluorescence images were obtained using a laser scanning confocal microscope (FV10i, Olympus), BX51 fluorescence microscope (Olympus) and ORCA-ER (Hamamatsu photonics). Time-lapse images were acquired using the FV10i-LIV. All figures were processed using Adobe Photoshop software.

Detection of apoptotic cells
Programmed cell death was detected by counting the COS-7 cells that expressed GFP or GFP-GAREM under a fluorescence microscope following Annexin VAlexa594 (Molecular Probes) staining. More than 200 GFP positive cells were detected. Three independent sets of experiments were conducted. Error bars show the standard deviation.

GAREM2 but not GAREM1 forms unique granular structures upon cell growth factor stimulation
To analyze the subcellular localization of each GAREM subtype in neuronal cells, immunofluorescence staining with specific antibodies against each GAREM subtype in NGFstimulated PC-12 cells was performed. We found that only the GAREM2 protein was aggregated in an NGF stimulation-dependent manner (Fig. 1b). This phenomenon was also confirmed in COS-7 cells expressing GFP fused-GAREM2 protein. In quiescent cells, there was no significant difference in the subcellular localization of GAREM1 and GAREM2 in the cytoplasm, which are predominantly localized to the cell membrane. However, in the present study, time-lapse analysis revealed a significant difference in the subcellular localization of each GAREM subtype in COS-7 cells after EGF stimulation. Whereas GFP-tagged GAREM1 remained near the actin-rich cell membrane, such as lamellipodia-like structures even after EGF stimulation for 15 min, GFP-tagged GAREM2 formed a dot-like granular structure after EGF stimulation ( Fig. 1c and Additional files 1 and 2). After long-term EGF stimulation, some GAREM2-expressed cells contained large granules with aggregated GAREM2 protein, but GAREM1-expressed cells did not (Fig. 1d). As shown in Fig. 1a, the two GAREM subtypes have several common functional regions; the only obvious difference between them was thought to be the presence (GAREM1) or absence (GAREM2) of a nuclear localization sequence. However, our results suggest that tyrosine-phosphorylated GAREM2, which is stimulated by cell growth factor, has a functional region not present in GAREM1 that is required for its characteristic cellular localization.

Identification of a region essential for EGF stimulation-dependent formation of GAREM2-specific granules
First, we genetically engineered five vectors to express GFP-fused chimeric GAREM1 and GAREM2 proteins to roughly identify the essential regions in GAREM2 (Fig. 2a). Three chimeric proteins, which were composed of the N-terminus of GAREM2 and the C-terminus of GAREM1, were localized as follows (Fig. 2b). Chimera1, which contained amino acids 1-110 of GAREM2 and the remainder of GAREM1, showed the same cytoplasmic localization as GAREM1. Since the N-terminus of GAREM2 has no NLS, its nuclear abundance was reduced compared to that of GAREM1. In contrast, Chimera2, which contained amino acids 1-330 of GAREM2 and the remainder of GAREM1, appeared to have GAREM2-specific granule localization. Furthermore, Chimera3, which contained amino acids 1-645 of GAREM2 and the remainder of GAREM1, showed nearly the same subcellular localization as GAREM2. The two other chimeric proteins contained the N-terminal sequence of GAREM1 and the C-terminus of GAREM2. The chimera composed of amino acids 1-110 of GAREM1 and the remainder of GAREM2 (Chimera4) showed granule formation. However, the localization of the final chimera (Chimera5), which contained amino acids 1-645 of GAREM1 and the remainder of GAREM2, was almost identical to that of GAREM1. The expression of each chimeric protein with the expected molecular weight was confirmed by immunoblotting with an anti-GFP antibody (Fig. 2c). These results suggest that the GAREM2-specific granules formed in COS-7 cells following EGF stimulation require amino acids 110-330 of GAREM2.

A short glycine-rich region in the N-terminal half of GAREM2 is required for granule formation
Based on the results of the above experiments, the N-terminal sequences of the two GAREM proteins in the region of amino acids 110-330 were compared in detail. Consequently, we focused on the hydrophobic amino acid-rich and glycine-rich regions (GRD) located around amino acids 170-210, which are only found in GAREM2 (Fig. 3a). Next, we generated and expressed two deletion mutants of GAREM2 that lacked this part of GAREM2 (Fig. 3b, c) and observed their subcellular localization. As expected, no EGF-stimulated granule formation was observed in the mutants lacking the GRD. In addition, newly emphasized here is that the granules of GAREM2 co-localized in part with the EGFR that migrated into the cell after ligand stimulation (Fig. 3d, e). This suggests that cell proliferation stimulus-dependent tyrosine phosphorylation of GAREM2 is essential for binding to EGFR and that the GRD might be involved not only in granule formation but also in the regulation of binding to EGFR.

Effect of the GRD in GAREM2 on its tyrosine phosphorylation and cell death induced with EGF stimulation
N-terminal FLAG-tagged GRD deletion mutants of GAREM2 were prepared for biochemical functional analysis of the GRD (Fig. 4a). COS-7 cells expressing these mutant proteins were stimulated with EGF, and GAREM2 was immunoprecipitated using a FLAG antibody. The results showed that GRD-deleted GAREM2 mutants had lower levels of tyrosine phosphorylation than the wild-type GAREM2 protein. The amount of bound EGF was reduced accordingly (Fig. 4b). These results suggest that the GRD may have regulated the tyrosine phosphorylation of GAREM2 and its binding to EGFR upon EGF stimulation. Furthermore, we found that cells overexpressing GFP-GAREM2 were more likely to undergo apoptosis upon EGF stimulation for 8 h (Fig. 4c). This phenomenon was suppressed in mutants of GAREM2, suggesting that protein aggregation of GAREM2 promotes cell death.

Discussion
In this study, we found that only brain-specific GAREM2 aggregates in a cell growth factor-stimulated manner. In addition, we found that the glycine-rich region of GAREM2 is essential for its aggregation and may affect its tyrosine phosphorylation and apoptosis. In the brain, several RNA-binding proteins with a glycine-rich region (TDP-43, FUS and TIA-1) are known to aggregate in stress granules when mutations are present and are involved in neurodegenerative diseases [19,20]. Compared to these RNA-binding proteins, the glycine-rich region of GAREM2 is shorter and has no sequence homology [16], Our data suggest that this glycine-rich region affects the conformational change of tyrosine-phosphorylated GAREM2 (Fig. 4). We also found that aggregated GAREM2 mostly does not co-localize with proteins in stress granules such as Staufen 1, G3BP or HuR (data not shown). However, as the exact physiological role of GAREM2 aggregation in the brain remains unclear, additional in-depth analyses are needed in the future. GAREM acts as an adaptor protein to transduce the signals of EGF or insulin family proteins, which contribute to brain function. Dysregulation of these signaling pathways in the brain is involved in neurodegenerative diseases such as Alzheimer's and Huntington's diseases [21,22]. In fact, ErbB family members are known to correlate with the pathological development of Alzheimer's disease [23,24].
Since mutations in the GAREM2 gene have not been found in patients with inherited neurodegenerative diseases, it is unlikely that GAREM2 is the direct causative gene for these diseases. However, GAREM2 has been newly identified as a positive effecter related to Alzheimer's disease and Huntington's disease by analyzing genome-wide gene expression data [11].
According to the results of a comprehensive behavioral battery, there were slight emotional differences between GAREM2 knockout and wild-type mice, and the knockout mice exhibited less anxiety, mildly increased social approach behavior, and longer latency to immobility. Additionally, neurite extension was suppressed in the primary culture of neurons derived from GAREM2 KO mice [10]. Studies using knockout animals have not revealed a relationship between GAREM2 function and neurodegenerative diseases. However, it is known that various proteins aggregate more in old mice [25], and this may be the case for GAREM2.
In addition, we expect that GAREM2 aggregation will be observed in cells that are highly stimulated by cell growth factors, such as tumor cells, in the brain. Therefore, in the future, we plan to observe the localization of GAREM proteins in various brain cells under such conditions. EGFR is known to regulate aging-related metabolic activity [26]. To study the relationship between GAREM2 and neurodegenerative diseases in individual mice, it is necessary to analyze the localization and function of GAREM2 not only in young mice but also in aged mice. We are currently developing a tyrosine phospho-specific GAREM2 antibody, which we hope will be a useful tool for detecting its aggregated form and might be a suitable marker to detect the progressing pathological condition of neuronal cells.
The results of this study may aid in the elucidation of GAREM2-specific functions and suggest that GAREM2 is a potential therapeutic target for neurodegenerative diseases.