Abstract
Estrogen accounts for several biological processes in the body; embryo implantation and pregnancy being one of the vital events. This manuscript aims to unearth the nuclear role of Son of sevenless1 (SOS1), its interaction with estrogen receptor alpha (ERα), and signal transducer and activator of transcription 3 (STAT3) in the uterine nucleus during embryo implantation. SOS1, a critical cytoplasmic linker between receptor tyrosine kinase and rat sarcoma virus signaling, translocates into the nucleus via its bipartite nuclear localization signal (NLS) during the ‘window of implantation’ in pregnant mice. SOS1 associates with chromatin, interacts with histones, and shows intrinsic histone acetyltransferase (HAT) activity specifically acetylating lysine 16 (K16) residue of histone H4. SOS1 is a coactivator of STAT3 and a co-repressor of ERα. SOS1 creates a partial mesenchymal–epithelial transition by acting as a transcriptional modulator. Finally, our phylogenetic tree reveals that the two bipartite NLS surface in reptiles and the second acetyl coenzymeA (CoA) (RDNGPG) important for HAT activity emerges in mammals. Thus, SOS1 has evolved into a moonlighting protein, the special class of multi-tasking proteins, by virtue of its newly identified nuclear functions in addition to its previously known cytoplasmic function.
Introduction
Receptor tyrosine kinase (RTK) and integrin pathways are the two key signaling pathways that coordinate the cellular activities during the process of blastocyst implantation (Mercurio 2002). In these pathways, the protein Son of sevenless (SOS) acts as the guanine nucleotide exchange factor (GEF) of Ras (Chardin et al. 1993). SOS (the Ras guanine nucleotide exchange factor, RasGEF) is a critical linker between RTK signaling and its downstream effector, the small GTPase, rat sarcoma virus (RAS). The homologues of the SOS gene designated mSOS1 and mSOS2 were identified by hybridizing the Drosophila gene to a mouse cDNA library under low stringent conditions (Bowtell et al. 1992). Although SOS1 and SOS2 have 70% similarity in their sequence, SOS1 is shown to have both long-term and short-term signaling while SOS2 is known to be capable of executing short-term signaling (Qian et al. 2000). The accepted model of SOS function is the RasGEF activity and recruitment of SOS to the plasma membrane via formation of SOS– Grb2 complex ultimately responsible for activation of the mature, membrane-bound Ras proteins (McCormick 1993, Pawson & Schlessingert 1993). The Sos GEFs can also be involved in Rac activation (Nimnual et al. 1998, Scita et al. 1999). Reports suggest the role of SOS1 in the proliferation of human renal cell carcinoma cells (Shinohara et al. 1997) and that components of the SOS–RAS–RAF–MAPK pathway, such as RASAL2 and SOS1 are downregulated during differentiation of human embryonic stem cells (hESCs) (Armstrong et al. 2006). SOS1 is associated with many diseases; gain-of-function mutations is prevalent in Noonan syndrome (Roberts et al. 2007); mutations particularly in various functional domains viz., Glu191Lys, Arg744Gly in N-terminal histone fold and CDC25 domains, respectively, in case of dilated cardiomyopathy (Cowan et al. 2020).
The identity of SOS signaling in the embryo implantation scenario comes from the evidence that targeted disruption of both alleles of mouse Sos1, which encodes a Ras-specific exchange factor, conferred mid-gestational embryonic lethality that was secondary to impaired placental development and was associated with very low placental extracellular signal-regulated kinase (ERK) activity. Sos1 is essential for intrauterine development, with homozygous null animals dying in mid-gestation in association with reported yolk sac and embryonic heart defects (Wang et al. 1997). SOS1 is under the regulation of estrogen as is evidenced by the presence of a putative ERE sequence ‘tGGgCATGTTGACCT’ in the promoter at −3457 position (Bourdeau et al. 2004). Reports also suggest that human epidermal growth factor receptor 2 upregulation resulted in estrogen receptor alpha (ERα) interaction with SOS1 and hyperstimulation of the mitogen-activated protein kinase ERK1/2 in breast cancer cells (Yang et al. 2004).
Armstrong et al. demonstrated that phosphorylated c-RAF and SOS1 are present at the cell membrane and within the cytoplasm and nucleus (Armstrong et al. 2006). The discovery of a classically known ‘cytoplasmic’ protein SOS1 at an unexpected locale, nucleus, provides an indication of the protein to possess functions not attributed to it previously. Embryo implantation being an estrogen-regulated process, we intend to evaluate the spatiotemporal distribution of SOS1 during the window of implantation, analyze whether there exists an association between ERα and SOS1, and understand the functional relevance of its nuclear presence. In this study, we demonstrate that SOS1 is an interacting partner of ERα in the nucleus during the ‘window of implantation’. A pertinent observation is the estrogen-mediated distinct nuclear localization of a known cytoplasmic molecule, SOS1 – specifically restricted to implantation sites – suggestive of its importance in implantation. Moreover, SOS1 is tyrosine phosphorylated in the nucleus during the ‘window of implantation’. Our computational analysis identified the presence of simple bipartite karyophilic clusters of arginine and lysine similar to nuclear localization signals (NLS) in nucleus-targeted protein signaling. We demonstrate estrogen-regulated nuclear migration of SOS1 via its bonafide NLS and its co-localization in the chromatin where it associates with histones H2A, H2B, and H4. SOS1 showed intrinsic histone acetyltransferase (HAT) activity – specifically acetylating histone H4. SOS1 associates with signal transducer and activator of transcription 3 (STAT3) and ERα in the nucleus during the late pre-implantation stage of pregnancy. We provide evidence that SOS1 is a coactivator of STAT3 on account of the presence of acetyl CoA binding motifs and a repressor of ERα by virtue of the presence of a CoRNR box with repressor motif L/V.x.x.I/V.I. Finally, we establish that SOS1 creates a partial MET with increased E-cadherin and cytokeratin expression albeit with increased invasion. Hence, SOS1 is a moonlighting protein with multiple functions of regulating STAT3 and ERα, and a HAT in addition to its known GEF functions.
Materials and methods
Sos 1 siRNA (h) (sc-29486) (NM_005633 with a mix of 3 siRNAs with sequence details as GACACAUAUUUCUCUUUGAtt, CAGCAUUGAUAUCUUUACAtt, CUCUCCUAAUCUUCUGAAAtt) was purchased from Santa Cruz Biotechnology while Biotrace™ PVDF (0.45 µm) was from PALL Corporation (Port Washington, NY, USA). Goat anti-rabbit FITC, propidium iodide, HEPES, progesterone, 17β-estradiol, corn oil, fetal bovine serum (FBS), insulin (I6634) and other chemicals were from Sigma Chemicals. BIO-RAD DC Protein Assay kit was from Bio-Rad Laboratories, while goat anti-rabbit Alexa Fluor 488 (A11070), DMEM-F12, DMEM, antibiotic–antimycotic cocktail, and trypsin-EDTA were from Invitrogen.
Experimental organism
Female Mus musculus (Swiss strain) bred in the institute animal facility and housed at a temperature of 27°C in light controlled rooms (14 h light:10 h darkness) were used for the study. This study was approved by RGCB Institutional Animal Ethical Committee vide approval # IAEC/114/MAL/2010 and IAEC/44/MAL/2006. Details of animal models used in the study are provided in Supplementary Experimental Procedures (see section on supplementary materials given at the end of this article).
Animal models
Pregnant mice model
Experiments were performed on regularly cycling sexually mature virgin female mice of the age group, 3–5 months and tissues were isolated on different days of ‘window of implantation’ as detailed elsewhere (Nautiyal et al. 2004). Briefly, the animals were put for mating and the day of vaginal plug was marked as Day 1 of pregnancy. The animals were euthanized on early pre-implantation (Day 4, 10:00 h), late pre-implantation (Day 4, 16:00 h), peri-implantation (Day 5, 05:00 h), and late peri-implantation (Day 5, 10:00 h) In addition, implantation and inter-implantation sites of the peri-implantation uteri were dissected out after i.v. injection of 1% Evans Blue dye via tail vein, 15 min prior to the sacrifice of animals and stored separately at −80°C until further experiments. All the experiments were done on three different animals (n = 3).
Delayed implantation model
To evaluate the regulation of steroid hormones viz. estrogen and progesterone, on the expression profile of SOS1, the delayed implantation model was employed. Pregnant females were bilaterally ovariectomized under ketamine (3 mg/25 g)–xylazine (0.4 mg/25 g) cocktail on the 3rd day of pregnancy after 16:00 h and were distributed into 2 batches A and B as detailed elsewhere (Nautiyal et al. 2004). The animals were euthanized at 05:00 h on day 8 of pregnancy and the tissues were processed for immunofluorescence studies and Western blotting analysis. All the experiments were done on three different animals (n = 3).
Cytosolic, nuclear, and membrane extract preparation
Cytosolic, nuclear, and membrane extracts of the uteri were prepared using CelLytic NuCLEAR Extraction Kit (Sigma) as per the manufacturer’s instructions with minor modifications as detailed later. In short, the tissues were minced and homogenized at 15,000 rpm (3 strokes/30 s) using a polytron homogenizer at 4°C in 1× lysis buffer. The homogenates were incubated for 60 min on ice on Genei Rocker-100 and centrifuged in Eppendorf 5820R centrifuge at 1000 g at 4°C for 45 min to pellet nuclei. The supernatants were aspirated and centrifuged at 10,000 g to pellet mitochondria. The post mitochondrial supernatants were subjected to ultracentrifugation at 100,000 g in Beckman L8-M ultracentrifuge using NVT 90 rotor to pellet membranes. The supernatants thus obtained were stored at −80°C as cytosolic extracts. The pellets were washed three times using hypotonic buffer at 4°C and resuspended in the same buffer and stored as membrane extracts. Nuclear proteins were extracted from the nuclear pellet by adding extraction buffer, and after vortexing, they were incubated on ice for 30 min on a rocker followed by centrifugation for 30 min at 20,000 g. The supernatants were stored as nuclear extracts at −80°C.
SDS-PAGE, Western blotting, and immunoreaction
Equal amounts of protein (30 µg) from the cytosolic or nuclear extracts of different stages of pregnancy were diluted 1:2 with Laemmli sample buffer and blotted onto PVDF membrane after resolving on SDS-PAGE gels as detailed elsewhere (Nautiyal et al. 2004). In brief, blots were developed with 1% BSA as a blocker and appropriate primary and secondary antibodies. SOS1 (H-122) (sc-10803), ERα (sc-542), Stat3 (sc-482), and goat anti-rabbit HRP (sc-2030) from Santa Cruz Biotechnology were used for the study. Primary antibody (200 µg/mL) at 1:500 dilution and secondary antibody at 1:5000 dilution were used for Western blot development. Images were captured by Bio-Rad Gel doc.
Immunocytochemistry on the pregnant and delayed uterine sections
The excised uteri were processed for sectioning and immunolabeling as described earlier (Nautiyal et al. 2004). Minor modifications included the use of 1% BSA as a blocker, overnight incubation with primary antibody (1:100) at 4°C, and incubation with FITC-conjugated secondary antibody (1:200) in dark for 60 min and used propidium iodide to stain the nuclei. For co-localization, the steps were repeated from blocking for the second protein. The sections were further processed for staining with 50 µg/mL solution of 4, 6-diamidino-2-phenylindole-dihydrochloride hydrate (DAPI) to detect the nuclei. Sections were imaged using a Leica SP2 confocal laser scanning microscope. All experiments were repeated at least three times with three different animals.
Cell culture
RL95-2 (CRL-1671, ATCC, USA), an adhesive human endometrial cell line that can mimic functional endometrium on hormone supplementation was used for the study and was maintained in the laboratory according to ATCC protocol. The cells were grown in 10−7 M progesterone background in ATCC prescribed media and were treated with water-soluble estrogen (10−8 M) (Lin et al. 2007) for 2 time points (i) 0–30 min and (ii) 0–72 h. The cells were then fixed using 4% paraformaldehyde, neutralized in 50 mM ammonium chloride, and permeabilized with 0.25% Triton X 100. The non-specific binding sites were blocked with 3% FBS, probed with primary antibody (1:200) and Alexa fluor 488 conjugated secondary antibody (1:400). Propidium iodide (1:10,000) was used to stain the nucleus. The coverslips were mounted on the slides using DPX. Images were captured using Leica SP2 confocal microscope. HEK 293 cells (ATCC) were maintained in the laboratory according to ATCC protocol. HEK cells have high transfection efficiency and lack functional ERα, and hence were used for the study. Prior to transfection, cells were grown in 35 mm culture dishes (In Vitro Scientific, Mountain View, CA, USA) incubated at 37°C in CO2 incubator after transfection.
Site-directed mutagenesis
mSOS1-EGFP plasmid (purchased from Genecopoeia, Rockville, MD, USA) was used as a template for site-directed mutagenesis. The basic residues in the NLS motif (KK727AA, RK739AA, RR949AA, and KR960AA) and co-repressor motif (V852A, II855AA) were mutated by alanine substitution and SOS1Δ744 EGFP was made by deleting the 18-bp sequence (RDNGPG) motif using GeneTailor Site-Directed Mutagenesis kit (Invitrogen) as per the manufacturer’s instructions. Primers for site-directed mutagenesis were made using PrimerX. PCR-generated templates and site-directed mutants were verified by DNA sequence analysis using Sanger’s dideoxy method and AmpliCycle Sequencing kit of Applied Biosystems.
Transfection of HEK293 cells
Lipofectamine (Invitrogen)-mediated transient co-transfection was done in HEK 293 cells to understand the subcellular expression pattern under an estrogen induction. mERα-DsRed constructs made earlier from uterine ERα (GENEBANK Accession numbers, EU791538; EU791540) and mSOS1-EGFP (purchased from Genecopoeia) were used for co-transfection. Transfection was performed with 1–2 μg of the plasmids using lipofectamine 2000 reagent from Invitrogen as per manufacturer’s instructions. DMEM along with 1× FBS and antibiotic–antimycotic cocktail (Gibco) was used till capturing the images (6–18 h). Time-lapse confocal imaging was done after the addition of 10−8 M 17β-estradiol.
Confocal imaging
Images were acquired using NIKON ECLIPSE Ti microscope with a Nikon A1R confocal unit. Cells were subjected to similar conditions of laser power and voltage and observed under Apo VC 60X Oil immersion objective (NA-1.40). Images were then integrated into the NIS elements software for further analysis and measurements. Archived images of 16 bits were analyzed for fluorescence intensity in HEK293 cells for the subcellular expression pattern of SOS1 before and after the treatment. Average intensity values for a region of interest in pixel (mean intensity × ROI area covering cytosol) were obtained. Each ROI was selected after subtraction of background fluorescence intensity values of cytoplasm for better visualization. Each experiment was run in triplicate.
Pull-down assay with transfected cells and immunoblotting
Total cell extracts were prepared by using low salt detergent buffer and pull down assays were performed as detailed elsewhere (Padmanabhan et al. 2011). Briefly, the cells were washed with PBS and resuspended in IP lysis buffer, 25 mM Tris–Hcl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% NP-40, 5% glycerol, and protease cocktail inhibitor (Santa Cruz). The cell lysate was prepared and precleared using Protein A agarose beads (Santa Cruz) for 1 h and centrifuged for 5 min at 9800 g at 4°C. The primary antibody (1 µg) was added to the precleared lysate, incubated overnight at 4°C and agarose beads (25 µL) were used to capture the immune complex. Mock immunoprecipitation with protein A agarose alone was done for all the immunoprecipitation experiments. The complex was centrifuged for 10 min at 9800 g at 4°C to collect as pellet. The pellet was washed three times with ice-cold PBST to completely remove the supernatant to reduce the background. The protein–bead complex was denatured using Laemmli sample buffer (Nautiyal et al. 2004), centrifuged and the supernatant was loaded onto a 10% SDS-PAGE and blotted onto PVDF membrane. Immunoblotting was performed on the membrane to confirm the interacting partners.
HAT activity assay using wild and mutant SOS1
To prove the HAT activity of SOS1, we performed the HAT activity assay using the HAT activity colorimetric assay kit from BioVision. ERα, SOS1, ERα-SOS1, and ERα- Δ744 SOS1 in combination were used for the assay. Briefly, the kit utilizes active nuclear extract as a positive control and acetyl-CoA as a cofactor. Acetylation of peptide substrate by active HAT releases the free form of CoA which then serves as an essential coenzyme for producing NADH. NADH can easily be detected spectrophotometrically upon reacting with a soluble tetrazolium dye. The sample is read at OD440nm.
In vitro acetylation assay to confirm the HAT specificity
Histone extraction
Cells were harvested and washed twice with ice-cold PBS. The cells were resuspended in Triton extraction buffer (TEB: PBS containing 0.5% Triton X 100 (v/v), 2 mM phenylmethylsulfonyl fluoride (PMSF), 0.02% (w/v) NaN3) at a cell density of 107 cells/mL and lysed on ice for 10 min with gentle stirring. The samples were then centrifuged at 1500 g for 10 min at 4°C to pellet the cells. The cells were then washed in half the volume of TEB and centrifuged as before. The pellet was resuspended in 0.2 N HCl at a cell density of 4 × 107cells per mL. The histones were acid extracted overnight at 4°C. The samples were centrifuged at 1500 g for 10 min at 4°C. The supernatant was removed and protein content was determined using the Bradford assay.
Non-isotopic in vitro HAT assay
Non-isotopic in vitro HAT assay was also done as per the protocol described elsewhere (Kuninger et al. 2007). Briefly, HEK cells were transfected with wild and mutant Δ744 SOS1 and immunoprecipitated with EGFP antibody. The immunoprecipitates were washed twice in phosphate-buffered saline containing 0.1% Tween 20 (PBS-T), and once in acetyltransferase assay buffer (50 mM Tris-Cl pH 8, 10% glycerol, 10 mM butyric acid, 0.1 mM EDTA, 1 mM DTT, 1 mM PMSF). Unless otherwise specified, individual reactions contained immunoprecipitated proteins from 50 µg of nuclear extract in 50 µL of assay buffer with 10 µM acetyl CoA and 5 µg of purified histones. Acetylation reactions were incubated for 45 min (unless otherwise specified) at 30°C on a rotating platform, followed by the addition of SDS-PAGE sample buffer, electrophoresis through 10% SDS-PAGE gels, and transfer to PVDF membranes. Proteins were detected by immunoblotting followed by image acquisition using acetylated lysine antibody.
HAT activity assay
HAT activity assay with purified SOS1: to prove the HAT activity of SOS1, we performed HAT activity assay using the HAT activity colorimetric assay kit from BioVision. SOS1 GEF domain was purchased from Cytoskeleton (Denver, CO, USA); GRB2, sFRP4, and BSA were also used for the assay. Briefly, the kit utilizes active nuclear extract as a positive control and acetyl-CoA as a cofactor. Acetylation of peptide substrate by active HAT releases the free form of CoA which then serves as an essential coenzyme for producing NADH. NADH can easily be detected spectrophotometrically upon reacting with a soluble tetrazolium dye. The sample is read at OD440nm. Each experiment was run with a minimum of three replicates.
Ras activation assay
Since the HAT domain of SOS1 was in the CDC25 region, the RAS activity of SOS1 was assayed to study the impact of HAT domain mutation on RAS activity. GLISA Ras activation assay Biochem Kit was used for the assay. Each experiment was run in triplicates.
Real-time analysis
Total RNA was extracted using Trizol reagent (Sigma); the samples were quantified using ND-1000 spectrophotometer (Nano-Drop Technologies, Wilmington, DE, USA). A total of 100 ng of total RNA was reverse transcribed to cDNA using Superscript® VILO cDNA Synthesis Kit and real-time analysis was done with Power SYBR®Green PCR Master Mix (Life Technologies). The relative expression was calculated using ∆∆Ct method with 18S rRNA as the endogenous control. The run was performed on 7900HT Fast Real-Time PCR system (Applied Biosystems) under standard cycling conditions. Melting curve analysis was done to ensure the product specificity. Each experiment was run in triplicate with three different biological samples.
Invasion assay
Invasion assay with SOS1-EGFP, empty EGFP, and control cells were done using BD BioCoat Tumor invasion assay system containing the BD Falcon FluoroBlok 24-multiwell insert as per manufacturer’s instructions. The fluorescence was measured using the Infinite 200 Tecan microplate reader. Each experiment was run with a minimum of three replicates.
Bioinformatic tools employed in the study
The bioinformatics tools employed in the study are LOCATE, PSORTII prediction (https://psort.hgc.jp/form2.html), and SubLoc (http://www.bioinfo.tsinghua.edu.cn/SubLoc) to predict the subcellular localization of the protein.
Mass spectrometry
We performed co-immunoprecipitation with Importin α 1/6 antibody using the total and subcellular extracts of co-transfected HEK 293 cells. Pulled-down product was gel resolved, silver stained, and bands were trypsin digested and analyzed using mass spectrometry.
Statistical analysis
All the SDS-PAGE and Western blotting experiments were repeated at least three times from three biological replicates. Western blots were digitized on Bio-Rad FluorS MultiImager. Band densities were determined using Phoretix 1-D image analysis software (version 5.20, Nonlinear Dynamics, Durham, NC, USA).
Data availability
All data needed to evaluate the conclusions in the paper are present in the paper or in the Supplementary Information.
Results
ERα and SOS1 interact in the nucleus
Immunoprecipitation with ERα using nuclear extract of peri-implantation period pulled down SOS1 as an interacting partner
Immunoprecipitation with ERα using nuclear extract followed by peptide mass fingerprinting of the bands obtained identified SOS1 as an interacting partner during peri-implantation period in pregnant mice (Fig. 1A and B). The association of SOS1 was further confirmed by the presence of immunopositive bands of SOS1 on Day 4, 16:00 h and Day 5, 05:00 h in ERα immunoprecipitates (Fig. 1A). Mock immunoprecipitations with protein A agarose alone did not show bands in SOS1 and ERα lanes. Immunoprecipitation with ERα during different days of pregnancy authenticated their interaction on early pre-implantation (Day 4, 10:00 h) and peri-implantation (Day 5, 05:00 h) periods – interaction being stronger during early pre-implantation (Fig. 1C). The lower panel in Fig. 1C shows the input protein level (histone H3) taken for immunoprecipitation.
ERα interacts with SOS1 in vivo and in vitro. (A) Western blot of ERα immunoprecipitate using nuclear extract of the late pre-implantation period (D4, 16:00 h) and peri-implantation period (D5, 05:00 h) probed with SOS1 (sc-10803) and ERα (sc-542). The lower blot in the panel shows input. The first lane shows mock. (B) Table showing the peptide mass fingerprinting data analyzed in MASCOT software. GB represents the gel band number, the second column is the matched number of peptides, and the third column indicates the score. The fourth column is the molecular weight, the fifth column represents sequence coverage, and the last is the assigned protein. (C) Western blot of SOS1 on ERα immunoprecipitate of different days of pregnancy using the SOS1 (sc-10803) antibody. The lower panel shows the input protein level. (D) Co-localization of ERα and SOS1 in the uterine section of peri-implantation (D5, 05:00 h) stage. ERα probed with the primary antibody (sc-542) was labeled with anti-rabbit secondary Alexa Fluor 488 and SOS1 probed with primary (sc-10803) was labeled with Alexa Fluor 568. DAPI was used as the nuclear stain. Co-localization is clear from the yellow color in the overlay of SOS1 and ERα. (E) Co-localization of ERα and SOS1 in RL95-2 cell line. (F) Western blots of ERα and SOS1 using the RL95-2 total cell extracts before and after estrogen treatment. Actin is used as loading control. See also Supplementary Fig. 1.
Citation: Journal of Molecular Endocrinology 70, 1; 10.1530/JME-22-0089
To establish the interaction of ERα with SOS1, co-localization studies were done in uterine sections during the ‘window of implantation’ (Supplementary Fig. 1). SOS1 expression is paralleled with its co-localization with ERα at Day 4, 10:00 h and Day 5, 05:00 h (Fig. 1D). The DAPI image completely superimposed with that of ERα and SOS1 thereby signifying that both the molecules get limited to the nucleus. To confirm that the expression of SOS1 is increased during the implantation window, real-time analysis of Sos1 expression was done in non-pregnant and pre-implantation stage. There is a significant increase in the expression of SOS1 in pre-implantation stage uteri compared to non-pregnant mice (P < 0.008) (Supplementary Fig. 1B). The immunohistochemistry of non-pregnant uteri also shows a scattered expression of SOS1 in the endometrium (Supplementary Fig. 1C).
ERα and SOS1 colocalize in RL95-2 cell lines
Co-localization of ERα-SOS1 in RL95-2 confirmed the interaction in the endometrium. At 0 and 0.5 h of E2 treatment, both ERα and SOS1 showed wide distribution in cytoplasm and membrane (Fig. 1E). However, after 1 h of E2 treatment, increased expression of both SOS1 and ERα was observed, along with their co-localization in the nucleus (Fig. 1F). Western blotting with ERα and SOS1 using total extracts made from estrogen-treated/untreated RL95-2 cells corroborated with the immunocytochemistry results (Fig. 1F).
SOS1 is nuclear in mouse uterine cells at implantation
Expression profiling of SOS1 in the uterus during ‘window of implantation’
A very prominent 170 kDa immunopositive band for SOS1 was seen in the cytosol as well as in the nucleus (Fig. 2A and B) in uterine extracts during the ‘window of implantation’. Relatively high levels of SOS1 are maintained in the cytosol of Day 4, 10:00 hafter which they show a reduction from Day 4, 16:00 h to Day 5, 10:00 h. The nuclear expression of SOS1 is discernible with the highest levels at Day 5, 05:00 h (Fig. 2B(i)). SOS1 levels in membrane extracts showed a reduction at Day 5, 05:00 h with a major fall in expression at Day 5, 10:00 h (Fig. 2C).
Estrogen triggers nuclear translocation of SOS1. (A, B and C) Western blot analysis of uterine mSOS1 expression (cytosolic/nuclear/membrane) at the ‘window of implantation’. (A) Image showing mSOS1 levels in the uterine cytosolic extracts, β-actin level in cytosolic extracts, (B) mSOS1 levels in nuclear extracts, histone (H2B) levels in nuclear extracts, (C) mSOS1 levels in membrane extracts, reference band from the Coomassie-stained SDS-PAGE gel of membrane extract. (D) mSOS1 was stained with Alexa Fluor 488, nuclei stained with propidium iodide and merged images are shown on the right. mSOS1 localization at the nonreceptive stage (D4, 10:00 h), (E) pre-implantation stage (D4, 04:00 h), (F) implantation site of peri-implantation stage (D5, 05:00 h – IMP), (G) inter-implantation site of peri-implantation stage (D5, 05:00 h IIMP) and (H) post-implantation stage (D5, 10:00 h). Western blot analysis of uterine mSOS1 expression in delayed implantation model (I) mSOS1 levels in total extracts, (K) mSOS1 levels in cytosolic extracts, (M) mSOS1 levels in nuclear extracts. (J) Histogram showing uterine total, (L) cytosolic and (N) nuclear levels of mSOS1 in the delayed implantation model. Lane 1. Marker; 2. Progesterone treated and 3. Progesterone and estrogen treated. β-actin and histone H2B served as the loading controls for total, cytosolic, and nuclear extracts, respectively. Panel in I, K, and M are made by taking lanes of one biological replicate. All the biological replicates are shown in Supplementary Fig. 2 and the shortlisted lanes are marked with an asterisk (*). See also Supplementary Fig. 2. (O and P) Immunolocalization of mSOS1 in the uterus of the delayed implantation model: mSOS1was stained with Alexa Fluor 488, nuclei stained with propidium iodide and merged images are shown on the right. (O) mSOS1 localization in the uterus of P alone model. (P) mSOS1 localization in uterus of P + E model.
Citation: Journal of Molecular Endocrinology 70, 1; 10.1530/JME-22-0089
Immunolocalization of SOS1 in pregnant uterine sections reveals its nuclear presence
Distinct green fluorescence in overlay images at Day 4, 10:00 h and Day 4, 16:00 h reveals SOS1 expression in the cytoplasm and membrane (Fig. 2D and E). In the implantation site of the Day 5, 05:00 h of the uterus, the molecule was limited to the nucleus (Fig. 2F), while the cells of inter-implantation site are conspicuous by cytoplasmic/membrane localization of SOS1 (Fig. 2G). On Day 5, 10:00 h, the molecule was located in the nucleus and cytoplasm (Fig. 2H).
Estrogen triggers the nuclear shuttling of SOS1 in the uterus of delayed implantation model
Western blotting analysis of delayed implantation uteri showed a significantly elevated expression of SOS1 immunopositive band at ~170 kDa position in P + E model in total extracts (P < 0.002) (Fig. 2I, J and Supplementary Fig. 2A), cytosolic extracts (P < 0.01) (Fig. 2K, L and Supplementary Fig. 2C), and nuclear extracts (P <0.01) (Fig. 2M, N and Supplementary Fig. 2E) when compared with respective P alone group.
Immunofluorescence analysis revealed that the molecule was located only in the extra-nuclear area in the animals which received progesterone alone (Fig. 2O). In the P + E model, SOS1 was located clearly in the nucleus as well as in the extra-nuclear region (Fig. 2P) arguing for an estrogen-regulated nuclear trafficking of SOS1 and co-localization.
SOS1 enters the nucleus via a classical bipartite NLS
Bioinformatic analysis predicts SOS1 to be nuclear
Our in silico analysis predicts SOS1 to be nuclear using LOCATE (Supplementary Fig. 3A), Subloc v1.0 (Supplementary Fig. 3B) with an accuracy of 84% and PSORT II with a 78.3% probability and an NLS score of 3.42 (Supplementary Fig. 3C). PSORT II analysis identified two bipartite NLS in the molecule at positions 727 (KKWVESITKIIQRKKIA) and at 949 (RRHGKELINFSKRRRVA).
Mutation in bipartite NLS of SOS1 affects its nuclear translocation ability
The importance of bipartite NLS residues in mediating nuclear translocation of SOS1 under estrogen response was assessed by mutating the underlined residues of NLS at 727 position KKWVESITKIIQRKKIA and at 949 RRHGKELINFSKRRRVA (Fig. 3A). Transient co-transfection with mERα-DsRed and mSOS1-EGFP in HEK 293 cell line was performed. Upon estrogen treatment (1 nM), a significant proportion of wild mSOS1-EGFP showed a nuclear translocation within a period of 24 h, in addition to cytoplasmic expression (Fig. 3B), while nuclear translocation of mutant mSOS1-EGFP with mutated NLS (both NL1 & NL2) was hampered (Fig. 3B and Video 1).
Importin α ferries SOS1 into the nucleus. (A) Schema showing SOS1 domains, location of bipartite NLS, and introduced mutation in the NLS region. (B) Co-transfection of wild and mutant SOS1-EGFP with mERα-DsRed with and without estrogen treatment in HEK cell line. See also Video 1. (C) Western blot of importin α immunoprecipitate probed with SOS1 and importin α. The lower panel shows the protein input. (D) Western blot of SOS1 immunoprecipitate probed with SOS1 and importin α during different days of pregnancy. The lower panel shows the protein input.
Citation: Journal of Molecular Endocrinology 70, 1; 10.1530/JME-22-0089
Video showing the hampering of nuclear translocation of mutant mSOS1-EGFP with mutated NLS (both NL1 & NL2) even after estrogen treatment. View Video 1 at https://player.vimeo.com/video/1069502765?h=318b0eb337.
Citation: Journal of Molecular Endocrinology 70, 1; 10.1530/JME-22-0089
The karyophilic cargo protein, importin α, mediates SOS1 nuclear entry
Mascot sequence analysis of peptide fragments of importin α pull-down products revealed the presence of SOS1 with importin α (Supplementary Fig. 4), suggesting that SOS1 interacts with it and enters the nucleus using the classical importin α pathway. This importin α-mediated SOS1 nuclear entry concept is further strengthened by our observation of presence of wild type SOS1 in importin α immunoprecipitates. The mutant SOS1 is not pulled down by importin α in co-immunoprecipitation studies as evidenced from lack of SOS1 immunopositive band (Fig. 3C), suggesting that the identified NLS is critical for importin α interaction. Presence of importin α in SOS1 immunoprecipitates in pregnant uterine cytosolic extracts (Fig. 3D) confirms that this interaction occurs endogenously.
SOS1 is present in chromatin
SOS1 was visible in chromatin of estrogen-treated RL95-2 cells grown on the background of progesterone (Fig. 4A). To confirm the chromatin binding, we isolated the chromatin and digested it with DNase I to release the chromatin-bound proteins. The immunopositive band in Western blot of SOS1 in DNase fraction confirmed the presence of SOS1 in chromatin. HEK 293 total extract served as a positive control for SOS1. Histone H2B served as a positive control for DNase I fraction (Fig. 4B).
SOS1 is present in chromatin, binds histones, and shows HAT activity. (A) Localization of SOS1 in chromatin in estrogen-treated cells grown on the background of progesterone. (B) Western blot of SOS1 in DNase fraction to confirm the presence of SOS1 in chromatin. HEK 293 total extract served as a positive control for SOS1 in the experiment. Histone H2B also served as a positive control for DNase I fraction. (C) Western blot of ERα IP probed with histone H2B and ERα during different days of pregnancy. (D) Western blot of SOS1 IP probed with SOS1 and histone H2B during different days of pregnancy. (E) Image shows the protein input panel for both C and D. (F) (i) Western blot of SOS1 IP in estrogen-treated RL 95-2 extracts probed with various histones viz., H1, H2A, H2B, H3, and H4. (ii) Histogram showing the differential interaction of SOS1 with histones. (G) Graph representing HAT activity assay using the HAT activity colorimetric assay kit from BioVision. The nuclear extract from transfected cells viz., ERα, SOS1 and ERα-SOS1 in combination was used for the assay. (H) HAT activity colorimetric assay with recombinant SOS1 GEF domain, GRB2, sFRP4. BSA served as a negative control. (I) Non-isotopic in vitro HAT assay using purified histones and immunoprecipitated SOS1 from SOS1, SOS1∆744, control transfected cells. Serum was used as the negative control for the experiment. Histones and SOS1-EGFP were shown as input. (J) Non-isotopic in vitro HAT assay using purified histones and immunoprecipitated SOS1 from SOS1, SOS1∆744, control transfected cells probed with specific acetylated histone H4 antibodies (K5, K8, K12, K16, and K20) to analyze which lysine residue is getting acetylated. (K) In vitro HAT assay using purified histones and recombinant SOS1GEF domain purified protein encompassing the HAT domain to confirm K16 acetylation activity of SOS1. GRB2 served as negative control and H4 shows the protein load. (L) (i) Co-transfection experiment with mERα-DsRed and SOS1-EGFP before and after estrogen treatment. (ii) Co-transfection experiment with mERα-DsRed and Δ744 SOS1-EGFP. Estrogen trigger did not stimulate the nuclear entry of SOS1∆744 EGFP which shows that the domain is important for nuclear translocation also. (M) RAS activity assay in cytoplasmic and nuclear extract from SOS1, Δ744 SOS1 and vector alone transfected cells. Nuclear extracts of both did not show any RAS activity confirming the fact that HAT activity and GEF activity are independent of each other.
Citation: Journal of Molecular Endocrinology 70, 1; 10.1530/JME-22-0089
SOS1 binds to different histones
The presence of SOS1 in chromatin prompted us to check the histone binding ability of SOS1 and ERα. Immunoprecipitation experiments with ERα and SOS1 using the nuclear extracts of different days of pregnancy pulled down H2B as one of the interacting partners. H2B interacted with ERα during Day 4, 16:00 h and Day 5, 05:00 h (Fig. 4C), while SOS1 interacted with H2B during Day 4, 10:00 h and Day 5, 05:00 h (Fig. 4D). Figure 4E shows the input panel.
The chromatin presence of SOS1 and its interaction with H2B propelled us to evaluate the SOS1 association with other histones. Our SOS1 IP experiments in estrogen-treated RL 95-2 extracts revealed that SOS1 interacts robustly with histones H2A, H2B, and H4 and weakly with H1 (Fig. 4F(i)). The blot is quantified to show the varying level of interaction (Fig. 4F(ii)). There was no interaction of SOS1 with histone H3.
Acetyl CoA binding site at 744 in SOS1 is accountable for HAT activity for SOS1
The presence of SOS1 in the chromatin and its association with histones led us to check for its HAT activity potential. While screening SOS1 sequence for HAT-related domains, SOS1 did not have any structural features of characteristic HATs like bromodomain/chromodomain but was found to contain acetyl CoA recognition/binding motifs (Arg/Gln-X-X-Gly-X-Gly/Ala) at 325–330 (QSIGEG) and 744–749 positions (RDNGPG) similar to that of GNAT and MYST. The alignment of SOS1 based on the property of amino acids with known HATs shows similarity in the region of acetyl CoA site at 744 (Supplementary Fig. 5). The nuclear extract from transfected cells viz., ERα, SOS1, and ERα-SOS1 with wild SOS1 and SOS1 deletion mutant (acetyl CoA binding site at 744 deleted) in combination was used for HAT activity assay. After normalizing with control (vector alone), ERα did not show any HAT activity but SOS1 alone (P < 0.003) and ERα–SOS1 combination (P < 0.02) showed significant HAT activity (Fig. 4G). There was a significant reduction in HAT activity in ER–SOS1∆744 (P < 0.02, P < 0.03) samples compared to ER–SOS1 and SOS1 alone, respectively. Recombinant SOS1 GEF domain showed HAT activity similar to that of the SOS1 transfected nuclear extract while GRB2, sFRP4 did not show any activity. BSA served as negative control (Fig. 4H).
Purified histones were treated with immunoprecipitated SOS1 from SOS1, SOS1∆744, control transfected cells as per the protocol for non-isotopic in vitro HAT assay. Serum was used as the negative control. The blot probed with acetylated lysine showed an immunopositive band specifically at histone H4 which confirms the fact that SOS1 can specifically acetylate histone H4 (Fig. 4I). The experiment was repeated with specific acetylated histone H4 antibodies (K5, K8, K12, K16, and K20) to analyze which lysine residue is getting acetylated. SOS1 specifically acetylates K16 (Fig. 4J). To confirm that HAT activity resides in SOS1, Recombinant SOS1GEF domain purified protein encompassing the HAT domain was purchased from Cytoskeleton and the experiment was repeated. The blot clearly showed K16 acetylation reconfirming the HAT activity of SOS1 (Fig. 4K). GRB2 served as negative control and H4 shows the protein load. An estrogen trigger did not stimulate the nuclear entry of SOS1∆744 EGFP, which shows that the domain is important for nuclear translocation (Fig. 4L(ii)) while the wild SOS1-EGFP translocated to the nucleus after estrogen treatment (Fig. 4L(i)). It is important to note that this domain is immediately preceded by the first NLS at 727.
To confirm that HAT activity is independent of RAS activity, the RAS activity assay was performed in SOS1 transfected cells. Cytoplasmic and nuclear extract from both SOS1 transfected and Δ744 SOS1 transfected cells were analyzed for RAS activity. An increase in Ras activity is seen in wild SOS1 transfected cells when compared to the control (P < 0.045). Cytoplasmic extracts from mutant showed increased RAS activity when compared to wild (P < 0.038) which could be attributed to accumulated cytoplasmic SOS1 presence due to its inability to undergo nuclear translocation (Fig. 4M). The vector alone showed no significant change in RAS activity (P < 0.36) confirming that SOS1 is solely responsible for increased RAS activity. Nuclear extracts of both did not show any RAS activity confirming the fact that HAT activity and GEF activity are independent of each other.
Immunoprecipitation with SOS1 in the uterine nuclear extract of late pre-implantation period pulled down STAT3 and ERα as interacting partners
Immunoprecipitation with SOS1 using nuclear extract followed by peptide mass fingerprinting of the bands obtained identified STAT3 and ERα as interacting partners during Day 4, 16:00 h (Fig. 5A and B). The presence of SOS1 was further confirmed by Western blotting of SOS1 in STAT3 immunoprecipitate during different days of pregnancy. Immunoprecipitation with STAT3 during different days of pregnancy authenticated their interaction on Day 4, 16:00 h. The lower panel in Fig. 5C shows the input protein level.
SOS1 is a co-activator of STAT3. (A) Immunoprecipitation with SOS1 antibody (sc-10803) using the nuclear extract of the late pre-implantation period. A total of 11 bands were obtained marked 1–11. GB1, GB 3, and GB 4 correspond to SOS1, STAT3, and ERα, respectively. (B) Peptide mass fingerprinting data analyzed in MASCOT software. (C) Western blot of SOS1 and STAT3 in STAT3 immunoprecipitate during different days of pregnancy. The lower panel in C shows the input protein level. (D) Western blot of SOS IP probed with p-tyrosine antibody during different days of ‘window of implantation’. (E) Western blot of GRB2 IP in sodium orthovanadate (inhibitor of tyrosine phosphatase activity)-treated cells probed with SOS1 and GRB2. Actin was used as the loading control. (F) Western blot of SOS1-IP using sodium orthovanadate-treated cells probed with SOS1, STAT3, ERα, and GRB2. (G) STAT3 reporter assays with wild and mutant SOS1 (SOS1∆744) transfected cells. (H) Real-time analysis of the downstream genes of STAT3 in wild and mutant SOS1 (SOS1∆744) transfected cells.
Citation: Journal of Molecular Endocrinology 70, 1; 10.1530/JME-22-0089
Tyrosine phosphorylation of SOS1 is crucial for its interaction with STAT3
SOS1 is tyrosine phosphorylated in the uterus
SOS is known to be phosphorylated causing uncoupling of SOS-Grb2 complex (Waters et al. 1995b) and was observed as a doublet in the uterine nuclear fraction, and we checked the phosphorylation status of SOS1. Western blot of SOS1 IP probed with p-tyrosine antibody showed an increasing degree of phosphorylation of SOS1 from Day 4, 10:00 h to Day 5, 05:00 h with maximal phosphorylation evident at Day 5, 05:00 h. A reduction in phosphorylation state is seen at Day 5, 10:00 h (Fig. 5D). Phosphorylated SOS1 normally dissociates from GRB2 complex; thus, SOS1 tyrosine phosphorylation could be crucial for its nuclear trafficking. GRB2 IP in sodium orthovanadate-treated cells (tyrosine phosphatase activity inhibitor) showed reduced interaction of GRB2 with SOS1 compared to the untreated cells confirming that tyrosine phosphorylated SOS1 dissociates from GRB2 complex (Fig. 5E).
Tyrosine-phosphorylated SOS1 interacts with STAT3
To evaluate whether the tyrosine phosphorylation of SOS1 is essential for its interaction with STAT3 and ERα, we performed SOS1 immunoprecipitation in sodium orthovanadate-treated cells and probed for the presence of STAT3 and ERα. Western blot of STAT3 in SOS1-IP using sodium orthovanadate-treated cells showed increased STAT3 association in treated cells, but the association with ERα remained the same (Fig. 5F), suggesting that phosphorylation of SOS1 is important for its interaction with STAT3. Western blot of GRB2 in SOS1 IP showed reduced association in sodium orthovanadate-treated cells compared to the untreated (Fig. 5F) which reaffirms the GRB2 IP data.
SOS1 is a co-activator of STAT3 function in the nucleus during ‘window of implantation’
As we found STAT3 and ERα in the immunoprecipitate, we screened for the regulatory potential of SOS1 in controlling both these molecules. STAT3 reporter assays with wild and mutant SOS1 (SOS1∆744) proved SOS1 to be a coactivator for STAT3 since cells transfected with wild SOS1 showed a 2.3-fold increase in the STAT3 promoter activity (P < 0.00004). Mutant SOS1∆744 showed a 6-fold reduction in STAT3 transactivation potential (P < 0.0001) (Fig. 5G). Appropriate controls viz., transfection with empty vectors show basal luciferase activity. Real-time analysis of the downstream genes of STAT3 also validates the hypothesis of a co-activator. SOS1 overexpression caused significant upregulation of STAT3 downstream genes viz., Bcl2 (P < 0.03), Hsp70 (P < 0.01),and Vegf (P < 0.00004) when compared to the control samples (Fig. 5H).
SOS1 is a repressor of ER function in the nucleus during ‘window of implantation’
We have identified two LXXLL motifs (NR-box), two inverted LXXLL motifs, and a co-repressor motif (V.S.R.I.I) in SOS1 (Fig. 6A), and this suggests that it could modulate ERE transactivation potential. ERE reporter assay with the wild SOS1 repressed ERE activity compared to the control cells. Since SOS1 was found to be a negative regulator of ERα, we mutated the co-repressor motif to analyze the importance of this motif in ER–SOS1 interaction. With estrogen treatment, there was a fourfold reduction in ERE transactivation potential (P < 0.004) in ER–SOS1 cotransfected cells (Fig. 6B). In cells co-transfected with ERα and the repressor mutants V852A (P < 0.00017) and II855AA mSOS1 (P < 0.00001), there was a significant increase in ERE transactivation potential to almost 5-fold proving the premise of it being a co-repressor. Thus, ERE reporter assays with wild and mutant SOS1 (V852A/ II855AA mSOS1) along with ERα proved SOS1 to be a co-repressor.
SOS1 is a co-repressor of ERα. (A) Alignment showing two LXXLL motifs, two inverted LXXLL motifs, and a co-repressor motif in SOS1. (B) ERE reporter assays with wild and mutant SOS1 (V852A and II855AA mSOS1) transfected cells. (C) Co-transfection of SOS1 both wild and repressor mutants (V852A and II855AA mSOS1-EGFP) with mERα-DsRed in HEK cell line to check their interaction capability and estrogen response. (D) Real-time analysis of the downstream genes of ERα in Sos1-silenced cells. (E) Real-time analysis of the proliferation markers in SOS1-silenced cells.
Citation: Journal of Molecular Endocrinology 70, 1; 10.1530/JME-22-0089
The co-transfection studies showed wild type SOS1 expression was limited to the cytoplasm of the cell and ERα was found in the nucleus (Fig. 6C). Upon estrogen treatment (1 nM) for 24 h, SOS1 migrated to the nucleus and exhibited clear co-localization with ERα (Fig. 6D). Mutations of the co-repressor motifs V852A and II855AA completely affected the interaction. V852 mutant was present in the nucleus before estrogen treatment but was present only in the cytoplasm after estrogen treatment, while II855AA was completely absent in the nucleus before and after estrogen treatment. Differential spatiotemporal distribution of wild/mutated SOS1 advocates that the co-repressor motif in SOS1 is vital for the nuclear translocation of the molecule in the cell. The co-repressor action of SOS1 is further reconfirmed as ERα downstream genes genes, viz., c-jun (P < 0.007), c-myc (P < 0.02), Fbl (P < 0.006), and IPP5 (P < 0.02) were significantly upregulated in the SOS1-silenced cells when compared to the control siRNA-treated samples under an estrogen stimulation (Fig. 6E).
SOS1 promotes a partial mesenchymal–epithelial transition in MCF-7 cells
The presence of SOS1 in the nucleus, its potential to activate STAT3, the penultimate controller of master transcription factors regulating EMT, made us investigate whether SOS1 can modulate EMT. Real-time analysis of the EMT-specific marker genes in SOS1-transfected cells showed significant increase in expression of E-cadherin (P < 0.01) and Cytokeratin (P < 0.004), while there was no change in the expression level of N-cadherin (P < 0.6767) and Vimentin (P < 0.39) (Fig. 7A and B).
SOS1 promotes a partial mesenchymal–epithelial transition in MCF-7 cells. (A and B) Real-time analysis of the EMT-specific marker genes in SOS1 transfected cells viz., E cadherin, Cytokeratin, N-cadherin, and Vimentin. (C) Real-time analysis of invasion markers MMP-2, MMP-9, and TIMP2. (Di) Graph representing Invasion assay with SOS1 and Δ744 SOS1 using the BD Biocoat tumor invasion assay kit. (Dii) Image representing the number of cells invaded in wild and mutant SOS1 transfected wells. (E) A loop of activation and repression involving SOS1, STAT3, and ERα in the process of epithelial–mesenchymal transition. (F) Cartoon representing the cytoplasmic and newly identified nuclear roles of SOS1.
Citation: Journal of Molecular Endocrinology 70, 1; 10.1530/JME-22-0089
The markers of invasion viz., Mmp-2 level (P < 0.027) was significantly decreased in SOS1 transfected cells, while Mmp-9 (P < 0.03) showed a significant increase (P < 0.03) (Fig. 7C). Timp2, the regulator of Mmp-9 was reduced in SOS1 transfected cells (P < 0.004). Mmp-9 is also downstream of STAT3 which could be the reason for its increase. Invasion assay also showed that SOS1 can increase invasion. The increase in Mmp-9 and decrease in Mmp-2 under estrogen trigger denotes the fact that SOS1 could be a good modulator of invasion (Fig. 7C and D).
Discussion
We demonstrate that SOS1 translocates to the nucleus via its NLS under an estrogen switch. It can bind to chromatin by interacting with histones. It executes its nuclear function via its HAT activity as it harbors an acetyl CoA binding site, thereby regulating transcription. Further, it works as a co-repressor of ERα and a coactivator of STAT3. Thus, it modulates several functions finally causing a partial MET and hence can multitask earning its place as a moonlighting protein.
Although one report documents SOS1 to be nuclear (Armstrong et al. 2006), its mechanism of nuclear entry remains elusive. Migration of cytoplasmic proteins viz., focal adhesion factors (including zyxin, paxillin, Abl-1) and receptor proteins (EGFR, Abl, ICAP) to the nucleus making use of their arginine-rich NLS (Hervy et al. 2006, Lo et al. 2006) has been reported although their nuclear function remains unclear. It is well established that NLS (Henderson 2000) and nuclear export signals (Tatematsu et al. 2005) help nuclear import and export. SOS1 nuclear entry is by virtue of its bipartite NLSs as NLS mutant constructs fail to shuttle into the nucleus reaffirming that the identified NLS is a valid one crucial for its nuclear translocation. The identification of importin α as an associating partner with SOS1 further establishes that SOS1 translocates to the nucleus by recognition of its NLS by importin α (Soniat & Chook 2015).
Delineation of the nuclear function of SOS1 comes from our finding of its presence in chromatin fraction under stimulation of estrogen and progesterone and its capability to interact with various histones specifically H2A, H2B, and H4. Acetyl-coenzyme A (CoA)-binding motifs are typical of HAT proteins (Sterner & Berger 2000). In-depth analysis of SOS1 sequence discovered acetyl CoA binding motif (Arg/Gln-X-X-Gly-X-Gly/Ala) which is responsible for HAT activity of SOS1 as mutation of acetyl-CoA site at 744 caused loss of HAT activity. Acetylation of K16 residue of histone H4 by SOS1 opens up a new role in modulating chromatin structure, inter-nucleosome interactions, gene expression, and DNA repair mechanisms (Zhang et al. 2017, Ray et al. 2018).
HAT activity is often associated with transcriptional activation. Our finding of SOS1 as a co-activator of STAT3 suggests that SOS1 with its HAT capability causes transcriptional activation of STAT3. We also find that phosphorylation of SOS1 is important for its interaction with STAT3 besides its already known relation with Grb2 dissociation. Sos-Grb2 disassociation is known to be caused by SOS1 phosphorylation via increased MEK/ERK due to Ras activation representing a feedback control loop to desensitize RAS activation (Waters et al. 1995a). Interaction of SOS1 with STAT3 opens up a new avenue for exploration of the role of SOS1 in mesenchymal–epithelial transition as STAT3 along with MCL-1 promotes MET (Renjini et al. 2014). We establish that SOS1 promotes a partial mesenchymal–epithelial transition with increased E-cadherin, cytokeratin expression and invasion. Sos1 silencing is known to cause a reduction in macrophage invasive capacity (Baruzzi et al. 2015). SOS1 is thought to be involved in podosome and invadopodia formation as it associates with tyrosine kinase substrate 5 (Rufer et al. 2009), a component of podosome and invadopodia (Buschman et al. 2009). Interestingly extravillous trophoblasts show atypical podosomes (Patel & Dash 2012) strengthening our observed role of SOS1 in invasion.
The expression and functions of molecules that are important in the process of implantation are tightly regulated by steroid hormones (Rueda et al. 2000, Salamonsen et al. 2002) through their receptors. Estrogen induces Shc phosphorylation and Shc-Grb2-Sos complex formation (Song et al. 2002). The strength of ERα signaling is modulated by steroid receptor interaction with coregulators. Our finding of ERα as an interacting partner of SOS1 in the nucleus and co-repressor potential of SOS1 suggests that it can rein in the transactivation capability of ERα as it harbors a co-repressor motif (V.S.R.I.I) similar to that present in NCoR and SMRT having CoRNR boxes with motif L/V.x.x.I/V.I. (Hu & Lazar 1999). Although SOS1 interaction with ERα in the cytoplasm is reported (Yang et al. 2004), our finding of ERα–SOS1 physical interaction in the nucleus adds a new dimension to the SOS1 saga.
Our study demonstrated an interaction between ERα–SOS1 and SOS1–STAT3. Reports are available regarding the interaction between ERα and STAT3 and that activated ERα blocks STAT3 transcriptional activity (Wang et al. 2001). Also, loss of ERα promotes EMT in breast cancer cells (Al Saleh et al. 2011) while ERα enhances the EMT program in the prostate epithelium (Shao et al. 2014). STAT3 or SOS1 alone can trigger EMT machinery as part of several signaling pathways (Gonzalez & Medici 2014, Wendt et al. 2014). Thus, we propose a loop of activation and repression involving SOS1, STAT3, and ERα which would control many processes of embryo implantation (Fig. 7E).
SOS1 is accounted in literature as a cytoplasmic molecule with Ras/Rac GEF activity critical in RTK/ cytokine/G-protein signaling. Here, we suggest new dimensions for SOS1 protein – a nuclear protein with a bipartite NLS, a coactivator of STAT3, a co-repressor of ERα, and a histone binding protein endorsed with HAT activity showing its capability to efficiently multitask and hence be a moonlighting protein (Fig. 7F).
SOS1 acquired moonlighting capability by sequential modification in its structure during metazoan evolution
When analyzing the sequence of available SOS1 homologs in NCBI, we found that several motifs were successively accepted by SOS1 during evolution. SOS1 was discovered two decades ago during the study of composed eye development in Drosophila as a GDP/GTP exchanger to RAS (Simon et al. 1991). Shortly, SOS1 presence was identified in Caenorhabditis elegans to mammalian cells (Bowtell et al. 1992, Chardin et al. 1993, Chang et al. 2000) and also the single gene evolved into two in mammals as SOS1 and SOS2. The domains involved in the GEF activity were conserved during the process while the domains that we identified started appearing from Drosophila onwards. The phylogentic tree (Cladogram) shows the relationship among different species on account of the SOS1 sequence (Fig. 8A). The tree was generated with SOS1 sequences of Clonorchis sinensis, Caenorhabditis elegans, Strongyloides ratti, Drosophila melanogaster, Danio rerio, Anolis carolinensis, Gallus gallus, Mus musculus, Rattus norvegicus, Bos taurus, Physeter catodon, Callithrix jacchus, Homo sapiens, and Pan troglodytes. Figure 8B, C and D represents the alignment of SOS1 sequences among these species highlighting the acetyl CoA motifs (in red) and NLS regions (in green). It is important to note that Drosophila with its first NLS region (Fig. 8C) and Danio rerio (zebrafish) with the second NLS region (Fig. 8D) have one NLS each. Acquisition of both the bipartite NLS appears to have occurred from reptile (Anolis carolinensis) onwards during the course of evolution. Interestingly, the bird Gallus gallus (chicken) and the reptile (Anolis carolinensis – green anole) do not possess the second acetyl coA (RDNGPG) which is important for HAT activity. Thus, it appears that HAT activity feature of SOS1, which participates in both short- and long-term signaling, evolved in mammals. Although SOS1 and SOS2 have 70% similarity in their sequence, SOS2 lacks the acetyl CoA binding motifs and SOS2 is known to be capable of executing short-term signaling (Qian et al. 2000). Gene duplication plays a leading role in origin and management of moonlighting proteins (Espinosa-Cantu et al. 2015) and in the case of SOS, gene duplication with acquisition of specific motifs during the course of evolution could have enabled it in separating the functions of short-term and long-term signaling.
Evolutionary modifications in SOS1 protein domains. (A) The phylogentic tree (Cladogram) showing the relationship among different species on account of the SOS1 sequence. The tree was generated with SOS1 sequences of Clonorchis sinensis, Caenorhabditis elegans, Strongyloides ratti, Drosophila melanogaster, Danio rerio, Anolis carolinensis, Gallus gallus, Mus musculus, Rattus norvegicus, Bos taurus, Physeter catodon, Callithrix jacchus, Homo sapiens, and Pan troglodytes. (B, C and D) represents the alignment of SOS1 sequences among these species highlighting the acetyl CoA motifs (in red) and NLS regions (in green). It is interesting to note that Drosophila with its first NLS region and Danio rerio (zebrafish) with the second NLS region have one NLS each.
Citation: Journal of Molecular Endocrinology 70, 1; 10.1530/JME-22-0089
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/JME-22-0089.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by BRNS grant no. 2010/37B/35/BRNS/1432 and in part by Department of Biotechnology – RGCB Core funds to Dr Malini Laloraya. Z D P received fellowship from CSIR (F. NO. 9/716(69)/2005-EMR-I), A P R was a recipient of a post doctoral fellowship from ICMR (3/1/2/14/12-RCH), India and A G S was supported by BRNS fellowship.
Author contribution statement
Project conception, design and supervision: M L. M L, Z D P, A P R, A G S and J N were involved in experimental designing and deliberations during the work. A P R contributed to Figs 1, 2A, B, C, 4C, E, F, G, H, I, J, K, L, M, 5B, C, D, E, F, G, H, I, 6, 7A, B, C, D, E, 8, Supplementary Figs 1, 5. Z D P was involved in Figs 2D, E, F, G, H, I, J, K, L, M, N, O, P, 4A, 5A, Supplementary Fig. 2, 3, and A G S contributed to Figs 3, 4D, Supplementary Fig. 4 and Video 1. M L and A P R contributed to Fig. 7F. M L, A P R and Z D P wrote the paper. All authors discussed the results and commented on the manuscript.
Acknowledgements
Ms. Jiji V, Bindu Asokan, Anurup K G, and Mr Manoj P are acknowledged for technical assistance in confocal imaging and sequencing.
References
Al Saleh S, Al Mulla F & Luqmani YA 2011 Estrogen receptor silencing induces epithelial to mesenchymal transition in human breast cancer cells. PLoS ONE 6 e20610. (https://doi.org/10.1371/journal.pone.0020610)
Armstrong L, Hughes O, Yung S, Hyslop L, Stewart R, Wappler I, Peters H, Walter T, Stojkovic P & Evans J et al.2006 The role of PI3K/AKT, MAPK/ERK and NFkappabeta signalling in the maintenance of human embryonic stem cell pluripotency and viability highlighted by transcriptional profiling and functional analysis. Human Molecular Genetics 15 1894–1913. (https://doi.org/10.1093/hmg/ddl112)
Baruzzi A, Remelli S, Lorenzetto E, Sega M, Chignola R & Berton G 2015 Sos1 regulates macrophage podosome assembly and macrophage invasive capacity. Journal of Immunology 195 4900–4912. (https://doi.org/10.4049/jimmunol.1500579)
Bourdeau V, Deschenes J, Metivier R, Nagai Y, Nguyen D, Bretschneider N, Gannon F, White JH & Mader S 2004 Genome-wide identification of high-affinity estrogen response elements in human and mouse. Molecular Endocrinology 18 1411–1427. (https://doi.org/10.1210/me.2003-0441)
Bowtell D, Fu P, Simon M & Senior P 1992 Identification of murine homologues of the Drosophila son of Sevenless gene: potential activators of ras. PNAS 89 6511–6515. (https://doi.org/10.1073/pnas.89.14.6511)
Buschman MD, Bromann PA, Cejudo-Martin P, Wen F, Pass I & Courtneidge SA 2009 The novel adaptor protein Tks4 (SH3PXD2B) is required for functional podosome formation. Molecular Biology of the Cell 20 1302–1311. (https://doi.org/10.1091/mbc.e08-09-0949)
Chang C, Hopper NA & Sternberg PW 2000 Caenorhabditis elegans SOS-1 is necessary for multiple RAS-mediated developmental signals. EMBO Journal 19 3283–3294. (https://doi.org/10.1093/emboj/19.13.3283)
Chardin P, Camonis JH, Gale NW, van Aelst L, Schlessinger J, Wigler MH & Bar-Sagi D 1993 Human Sos1: a guanine nucleotide exchange factor for Ras that binds to GRB2. Science 260 1338–1343. (https://doi.org/10.1126/science.8493579)
Cowan JR, Salyer L, Wright NT, Kinnamon DD, Amaya P, Jordan E, Bamshad MJ, Nickerson DA & Hershberger RE 2020 SOS1 gain-of-function variants in dilated cardiomyopathy. Circulation: Genomic and Precision Medicine 13 e002892. (https://doi.org/10.1161/CIRCGEN.119.002892)
Espinosa-Cantu A, Ascencio D, Barona-Gomez F & DeLuna A 2015 Gene duplication and the evolution of moonlighting proteins. Frontiers in Genetics 6 227. (https://doi.org/10.3389/fgene.2015.00227)
Gonzalez DM & Medici D 2014 Signaling mechanisms of the epithelial-mesenchymal transition. Science Signaling 7 re8. (https://doi.org/10.1126/scisignal.2005189)
Henderson BR 2000 Nuclear-cytoplasmic shuttling of APC regulates beta-catenin subcellular localization and turnover. Nature Cell Biology 2 653–660. (https://doi.org/10.1038/35023605)
Hervy M, Hoffman L & Beckerle MC 2006 From the membrane to the nucleus and back again: bifunctional focal adhesion proteins. Current Opinion in Cell Biology 18 524–532. (https://doi.org/10.1016/j.ceb.2006.08.006)
Hu X & Lazar MA 1999 The CoRNR motif controls the recruitment of corepressors by nuclear hormone receptors. Nature 402 93–96. (https://doi.org/10.1038/47069)
Kuninger D, Lundblad J, Semirale A & Rotwein P 2007 A non-isotopic in vitro assay for histone acetylation. Journal of Biotechnology 131 253–260. (https://doi.org/10.1016/j.jbiotec.2007.07.498)
Lin CY, Vega VB, Thomsen JS, Zhang T, Kong SL, Xie M, Chiu KP, Lipovich L, Barnett DH & Stossi F et al.2007 Whole-genome cartography of estrogen receptor alpha binding sites. PLoS Genetics 3 e87. (https://doi.org/10.1371/journal.pgen.0030087)
Lo HW, li-Seyed M, Wu Y, Bartholomeusz G, Hsu SC & Hung MC 2006 Nuclear-cytoplasmic transport of EGFR involves receptor endocytosis, importin beta1 and CRM1. Journal of Cellular Biochemistry 98 1570–1583. (https://doi.org/10.1002/jcb.20876)
McCormick F 1993 Signal transduction. How receptors turn Ras on. Nature 363 15–16. (https://doi.org/10.1038/363015a0)
Mercurio AM 2002 Lessons from the alpha2 integrin knockout mouse. American Journal of Pathology 161 3–6. (https://doi.org/10.1016/s0002-9440(1064149-1)
Nautiyal J, Kumar PG & Laloraya M 2004 17Beta-estradiol induces nuclear translocation of CrkL at the window of embryo implantation. Biochemical and Biophysical Research Communications 318 103–112. (https://doi.org/10.1016/j.bbrc.2004.04.005)
Nimnual AS, Yatsula BA & Bar-Sagi D 1998 Coupling of Ras and Rac guanosine triphosphatases through the Ras exchanger Sos. Science 279 560–563. (https://doi.org/10.1126/science.279.5350.560)
Padmanabhan RA, Nirmala L, Murali M & Laloraya M 2011 CrkL is a co-activator of estrogen receptor alpha that enhances tumorigenic potential in cancer. Molecular Endocrinology 25 1499–1512. (https://doi.org/10.1210/me.2011-0008)
Patel A & Dash PR 2012 Formation of atypical podosomes in extravillous trophoblasts regulates extracellular matrix degradation. European Journal of Cell Biology 91 171–179. (https://doi.org/10.1016/j.ejcb.2011.11.006)
Pawson T & Schlessingert J 1993 SH2 and SH3 domains. Current Biology 3 434–442. (https://doi.org/10.1016/0960-9822(9390350-w)
Qian X, Esteban L, Vass WC, Upadhyaya C, Papageorge AG, Yienger K, Ward JM, Lowy DR & Santos E 2000 The Sos1 and Sos2 Ras-specific exchange factors: differences in placental expression and signaling properties. EMBO Journal 19 642–654. (https://doi.org/10.1093/emboj/19.4.642)
Ray A, Khan P & Chaudhuri RN 2018 Regulated acetylation and deacetylation of H4 K16 is essential for efficient NER in Saccharomyces cerevisiae. DNA Repair 72 39–55. (https://doi.org/10.1016/j.dnarep.2018.09.009)
Renjini AP, Titus S, Narayan P, Murali M, Jha RK & Laloraya M 2014 STAT3 and MCL-1 associate to cause a mesenchymal epithelial transition. Journal of Cell Science 127 1738–1750. (https://doi.org/10.1242/jcs.138214)
Roberts AE, Araki T, Swanson KD, Montgomery KT, Schiripo TA, Joshi VA, Li L, Yassin Y, Tamburino AM & Neel BG et al.2007 Germline gain-of-function mutations in SOS1 cause Noonan syndrome. Nature Genetics 39 70–74. (https://doi.org/10.1038/ng1926)
Rueda BR, Hendry IR, Hendry WJ III, Stormshak F, Slayden OD & Davis JS 2000 Decreased progesterone levels and progesterone receptor antagonists promote apoptotic cell death in bovine luteal cells. Biology of Reproduction 62 269–276. (https://doi.org/10.1095/biolreprod62.2.269)
Rufer AC, Rumpf J, von Holleben M, Beer S, Rittinger K & Groemping Y 2009 Isoform-selective interaction of the adaptor protein Tks5/FISH with Sos1 and dynamins. Journal of Molecular Biology 390 939–950. (https://doi.org/10.1016/j.jmb.2009.05.025)
Salamonsen LA, Nie G & Findlay JK 2002 Newly identified endometrial genes of importance for implantation. Journal of Reproductive Immunology 53 215–225. (https://doi.org/10.1016/s0165-0378(0100087-0)
Scita G, Nordstrom J, Carbone R, Tenca P, Giardina G, Gutkind S, Bjarnegard M, Betsholtz C & Di Fiore PP 1999 EPS8 and E3B1 transduce signals from Ras to Rac. Nature 401 290–293. (https://doi.org/10.1038/45822)
Shao R, Shi J, Liu H, Shi X, Du X, Klocker H, Lee C, Zhu Y & Zhang J 2014 Epithelial-to-mesenchymal transition and estrogen receptor alpha mediated epithelial dedifferentiation mark the development of benign prostatic hyperplasia. Prostate 74 970–982. (https://doi.org/10.1002/pros.22814)
Shinohara N, Ogiso Y, Tanaka M, Sazawa A, Harabayashi T & Koyanagi T 1997 The significance of Ras guanine nucleotide exchange factor, son of Sevenless protein, in renal cell carcinoma cell lines. Journal of Urology 158 908–911. (https://doi.org/10.1097/00005392-199709000-00070)
Simon MA, Bowtell DD, Dodson GS, Laverty TR & Rubin GM 1991 Ras1 and a putative guanine nucleotide exchange factor perform crucial steps in signaling by the sevenless protein tyrosine kinase. Cell 67 701–716. (https://doi.org/10.1016/0092-8674(9190065-7)
Song RXD, McPherson RA, Adam L, Bao Y, Shupnik M, Kumar R & Santen RJ 2002 Linkage of rapid estrogen action to MAPK activation by ER {alpha}-Shc association and Shc pathway activation. Molecular Endocrinology 16 116–127. (https://doi.org/10.1210/mend.16.1.0748)
Soniat M & Chook YM 2015 Nuclear localization signals for four distinct karyopherin-beta nuclear import systems. Biochemical Journal 468 353–362. (https://doi.org/10.1042/BJ20150368)
Sterner DE & Berger SL 2000 Acetylation of histones and transcription-related factors. Microbiology and Molecular Biology Reviews 64 435–459. (https://doi.org/10.1128/MMBR.64.2.435-459.2000)
Tatematsu K, Yoshimoto N, Koyanagi T, Tokunaga C, Tachibana T, Yoneda Y, Yoshida M, Okajima T, Tanizawa K & Kuroda S 2005 Nuclear-cytoplasmic shuttling of a RING-IBR protein RBCK1 and its functional interaction with nuclear body proteins. Journal of Biological Chemistry 280 22937–22944. (https://doi.org/10.1074/jbc.M413476200)
Wang DZ, Hammond VE, Abud HE, Bertoncello I, McAvoy JW & Bowtell DD 1997 Mutation in Sos1 dominantly enhances a weak allele of the EGFR, demonstrating a requirement for Sos1 in EGFR signaling and development. Genes and Development 11 309–320. (https://doi.org/10.1101/gad.11.3.309)
Wang LH, Yang XY, Mihalic K, Xiao W, Li D & Farrar WL 2001 Activation of estrogen receptor blocks interleukin-6-inducible cell growth of human multiple myeloma involving molecular cross-talk between estrogen receptor and STAT3 mediated by co-regulator PIAS3. Journal of Biological Chemistry 276 31839–31844. (https://doi.org/10.1074/jbc.M105185200)
Waters SB, Holt KH, Ross SE, Syu LJ, Guan KL, Saltiel AR, Koretzky GA & Pessin JE 1995a Desensitization of Ras activation by a feedback disassociation of the SOS-Grb2 complex. Journal of Biological Chemistry 270 20883–20886. (https://doi.org/10.1074/jbc.270.36.20883)
Waters SB, Yamauchi K & Pessin JE 1995b Insulin-stimulated disassociation of the SOS-Grb2 complex. Molecular and Cellular Biology 15 2791–2799. (https://doi.org/10.1128/MCB.15.5.2791)
Wendt MK, Balanis N, Carlin CR & Schiemann WP 2014 STAT3 and epithelial-mesenchymal transitions in carcinomas. JAK-STAT 3 e28975. (https://doi.org/10.4161/jkst.28975)
Yang Z, Barnes CJ & Kumar R 2004 Human epidermal growth factor receptor 2 status modulates subcellular localization of and interaction with estrogen receptor alpha in breast cancer cells. Clinical Cancer Research 10 3621–3628. (https://doi.org/10.1158/1078-0432.CCR-0740-3)
Zhang R, Erler J & Langowski J 2017 Histone acetylation regulates chromatin accessibility: role of H4K16 in inter-nucleosome interaction. Biophysical Journal 112 450–459. (https://doi.org/10.1016/j.bpj.2016.11.015)
