G3BP1-linked mRNA partitioning supports selective protein synthesis in response to oxidative stress

Abstract

Cells limit energy-consuming mRNA translation during stress to maintain metabolic homeostasis. Sequestration of mRNAs by RNA binding proteins (RBPs) into RNA granules reduces their translation, but it remains unclear whether RBPs also function in partitioning of specific transcripts to polysomes (PSs) to guide selective translation and stress adaptation in cancer. To study transcript partitioning under cell stress, we catalogued mRNAs enriched in prostate carcinoma PC-3 cell PSs, as defined by polysome fractionation and RNA sequencing (RNAseq), and compared them to mRNAs complexed with the known SG-nucleator protein, G3BP1, as defined by spatially-restricted enzymatic tagging and RNAseq. By comparing these compartments before and after short-term arsenite-induced oxidative stress, we identified three major categories of transcripts, namely those that were G3BP1-associated and PS-depleted, G3BP1-dissociated and PS-enriched, and G3BP1-associated but also PS-enriched. Oxidative stress profoundly altered the partitioning of transcripts between these compartments. Under arsenite stress, G3BP1-associated and PS-depleted transcripts correlated with reduced expression of encoded mitochondrial proteins, PS-enriched transcripts that disassociated from G3BP1 encoded cell cycle and cytoprotective proteins whose expression increased, while transcripts that were both G3BP1-associated and PS-enriched encoded proteins involved in diverse stress response pathways. Therefore, G3BP1 guides transcript partitioning to reprogram mRNA translation and support stress adaptation.

Figure 1.

Isolation of PS-enriched and PS-depleted transcripts and their validation. (A) Detection of newly synthesized proteins in untreated or arsenite treated cells using AHA (azidohomoalanine)-mediated CLICK method (36). Cells cultured in AHA-free medium functioned as the control. Note that arsenite stress reduced the synthesis of new proteins. (B) Polysomal trace of untreated or arsenite treated cells. (C) Volcano plot showing PS- enriched or PS-depleted transcripts in response to arsenite treatment. Blue balls on the left depict PS-depleted transcripts (log₂fold −1.0 and below, P value < 0.05), with name of a few transcripts and corresponding pathways are shown in blue. Red balls on the right depict PS-enriched transcripts (log₂fold 1.0 and above; P value < 0.05), with name of a few transcripts and corresponding pathways are shown in red. Grey balls represent unchanged/non-significant transcripts. (D) Venn diagram comparing PSseq data and total transcriptome data. (E) Validation of PSseq data for selected transcripts by qRT-PCR using polysomal RNA extracted from vehicle-treated/untreated (UT) or arsenite-treated (ARS) cells. Mean values ± SD are shown for three independent experiments. ***P < 0.001; **P < 0.01; ns, non-significant. (F) Validation of PSseq data of selected transcripts using Western blotting of untreated/vehicle treated (UT) or arsenite-treated cells.

Figure 2.

Pathway analysis of PS-enriched and PS-depleted transcripts and functional studies. (A) Gene ontology analysis of PS-enriched and PS-depleted transcripts using Metascape software (http://metascape.org). Oxygen consumption rate (OCR) (B) and extracellular acidification rate (ECAR) (C) were determined through real-time measurements using the Seahorse XF24 Extracellular Flux Analyzer (see Methods for the details). Mean values ± SD are shown for five independent experiments. (D) Measurement of ATP in untreated or arsenite treated cells. Mean values ± SD are shown for three independent experiments. **P < 0.01. (E) Cell cycle analysis of untreated or arsenite treated cells stained with propidium iodide and analysed by FACS.

Figure 3.

APEX method for the extraction of G3BP1-associated proteins and transcripts. (A) Schema for extracting and comparing G3BP1-associated proteins and transcripts. G3BP1-APEX, represented by green dots, is present as diffused in the vehicle treated cells while it is present as both aggregated as well as diffused in the arsenite treated cells. See Methods for details. (B) APEX-G3BP1 colocalization by immunofluorescence was assessed using the SG protein, TIA-1. A part of the image of APEX-G3BP1 immunostaining from vehicle treated and arsenite treated cells is enlarged and shown at the bottom panels. Note that G3BP1 is mainly present as diffused in the vehicle treated cells (red arrows) while it is present as both aggregated (white arrows) and diffused (red arrows) in the arsenite treated cells. Scale, 10 um. (C) Western blot showing detection of biotinylated proteins (lanes 1–4 and lanes 6–9) as detected using anti-biotin antibodies (first upper panel), G3BP1-APEX fusion protein and G3BP1 (endogenous) as detected using anti-G3BP1 antibodies (second panel), G3BP1-APEX fusion protein and APEX alone as detected by anti-APX2 (APEX2) antibodies (third panel). GAPDH is used as the loading control. (D) Western blot analysis to detect proteins associated with APEX-G3BP1 complexes in unstressed and arsenite treated cells.

Figure 4.

G3BP1-associated proteins and pathway analysis. (A) Venn diagram showing a comparison and categorization (4 categories) of G3BP1-associated proteins in unstressed and arsenite stressed cells (see text for more details). (B) Proteins that come under the above different categories are listed in boxes with the corresponding colour shades. SG/G3BP1-interacting proteins, already reported in the literature are shown in white letters and new proteins identified in the current study are shown in black letters. (C) Gene ontology analysis using the above four categories of proteins using Metascape software (http://metascape.org).

Figure 5.

RNAseq of G3BP1 associated transcripts. Volcano plot showing G3BP1 associated transcripts with comparisons: (A) G3BP1-APEX-ARS/G3BP1-APEX-UT; (B) G3BP1-APEX-ARS/CTRL-APEX-ARS and (C) G3BP1-APEX-UT/CTRL-APEX-UT (see text for more details). (D) Venn diagram showing a comparison of G3BP1-associated transcripts in unstressed and arsenite stressed, dividing the transcripts into four groups, namely stress-dependent (green shaded), stress-sensitive (purple shaded), stress-independent (light-brown shaded), and transcripts with reduced or dissociated interactions (grey shaded) after arsenite stress.

Figure 6.

Validation of G3BP1-associated transcripts. Selected G3BP1-associated mRNAs extracted from vehicle treated or arsenite treated cells were subjected to qRT-PCR using primers specific to the transcripts as indicated in the figure. Mean values ± SD are shown for three independent experiments. ***P < 0.001; **P < 0.01; *P < 0.05.

Figure 7.

Compartmentalization of G3BP1-associated transcripts with PSs and their gene ontology analysis. (A) Venn diagram illustrating the compartmentalization of G3BP1-associated and dissociated/reduced transcripts within (PS-enriched) or away (PS-depleted) from PSs in stressed PC-3 cells revealed four categories; Cat. A-D. Gene ontology analysis of G3BP1 partitioned transcripts, (B) corresponding to Cat. A, (C) corresponding to Cat. B and (D) corresponding to Cat. C, using Metascape software.

Figure 8.

Protein expression and localisation of G3BP1-partitioned transcripts. (A–F) RNA in situ hybridization of different G3BP1-partitioned transcripts using mRNA probes as indicated (red panels). The in-situ slides were subjected to IF staining using anti-G3BP1 antibodies (green panels) to detect colocalization with G3BP1 in presence and absence of stress. A portion of each main figure panel is enlarged and shown as an insert. Scale, 10 um. (G) Western blot analysis of selected G3BP1-partitioned transcripts (see Materials and Methods for details). (H) Activated BAX detected using anti-BAX (2D2) antibodies in untreated or arsenite treated cells. A portion of each main figure panel is enlarged and shown as an insert. White arrows point to the spots representing activated BAX. Scale, 10 um. (I) Quantification of BAX activation. Mean values ± SD are shown for three independent experiments. **P < 0.01.