Tumor suppressor BTG1 promotes PRMT1-mediated ATF4 function in response to cellular stress

Abstract

Cancer cells are frequently exposed to physiological stress conditions such as hypoxia and nutrient limitation. Escape from stress-induced apoptosis is one of the mechanisms used by malignant cells to survive unfavorable conditions. B-cell Translocation Gene 1 (BTG1) is a tumor suppressor that is frequently deleted in acute lymphoblastic leukemia and recurrently mutated in diffuse large B cell lymphoma. Moreover, low BTG1 expression levels have been linked to poor outcome in several solid tumors. How loss of BTG1 function contributes to tumor progression is not well understood. Here, using Btg1 knockout mice, we demonstrate that loss of Btg1 provides a survival advantage to primary mouse embryonic fibroblasts (MEFs) under stress conditions. This pro-survival effect involves regulation of Activating Transcription Factor 4 (ATF4), a key mediator of cellular stress responses. We show that BTG1 interacts with ATF4 and positively modulates its activity by recruiting the protein arginine methyl transferase PRMT1 to methylate ATF4 on arginine residue 239. We further extend these findings to B-cell progenitors, by showing that loss of Btg1 expression enhances stress adaptation of mouse bone marrow-derived B cell progenitors. In conclusion, we have identified the BTG1/PRMT1 complex as a new modifier of ATF4 mediated stress responses.

Keywords: ATF4; BTG1; PRMT1; cellular stress; leukemia.

Figure 1. BTG1 affects cellular responses to stress

(A) Btg1 mRNA is upregulated in response to various stress stimuli. WT MEFs were treated with various stressors and mRNA levels of Btg1 and Btg2 relative to untreated cells are shown. Bars represent average data from four independent experiments ± SEM. (B) BTG1, but not BTG2 expression, is required for survival under cellular stress conditions. WT, Btg1^−/− and Btg2^−/− MEFs were challenged with ER stress-inducing drugs (thapsigargin, tunicamycin), glucose depletion, and amino acid limitation (glutamine starvation, Asparaginase (ASNase)). Metabolic activity was determined by MTS assay and cell survival relative to untreated cells (set at 100%) is shown. Bars represent average data from four independent experiments ± SEM. (C) Loss of BTG1 protects cells from stress-induced apoptosis. WT and Btg1^−/− MEFs were stressed with tunicamycin and apoptosis was measured by western blot for cleaved PARP protein. Bars represent average data from three independent experiments ± SEM. P-values are indicated with ***P < 0.001, **P < 0.01 and *P < 0.05 (two-tailed t-test).

Figure 2. Loss of BTG1 negatively affects ATF4-mediated gene expression

(A) Reduced expression of ATF4 target genes in Btg1^−/− cells upon glutamine depletion. WT and Btg1^−/− MEFs were stressed with glutamine starvation for 16 hours and qPCR was performed to measure the expression level of the seven ATF4 targets identified by gene expression analysis (Table 3). Data is presented as fold induction of mRNA (expression level of untreated samples were set to 1). Bars represent average data from four independent experiments ± SEM. P-values are indicated with ***P < 0.001, **P < 0.01 and *P < 0.05 (two-tailed paired t-test). (B) The expression level of Atf4 itself is not affected by the absence of Btg1. (C) The expression of ATF4 target genes in Btg1^−/− cells is also attenuated at protein level. Western blot shows the protein expression of two ATF4 targets following glutamine starvation. (D) ATF4 occupancy on target gene promoters is affected by loss of BTG1. WT and Btg1^−/− MEFs were stressed by glutamine starvation for 16 hours and ChIP was performed using an ATF4 antibody to determine ATF4 binding on the promoters of Ddit3, Atf3 and Fgf21. IP without antibody served as negative control. For each target gene, qPCR was performed using primers which recognize the ATF4 binding site (upstream) and a control region around 1.5 kb further (downstream). Bars represent average data from three independent experiments ± SEM. P-values are indicated with *P < 0.05 (two-tailed paired t-test).

Figure 3. BTG1 facilitates PRMT1 binding to and methylation of ATF4

(A) BTG1 binds to ATF4. HEK293 cells were transfected with expression plasmids encoding HA-ATF4 and FLAG-BTG1 and treated for 24 hrs with 5 μM of the proteasome inhibitor MG132. Protein lysates were generated and subjected to immunoprecipitation (IP) with FLAG antibody (Ab). Immunoblot demonstrates expression of BTG1 using a FLAG-Ab. (B) PRMT1 binds to ATF4 in a BTG1-dependent manner. WT and Btg1^−/− MEFs were treated with glutamine starvation for 16 h. Protein lysates were generated and subjected to IP with PRMT1 Ab. Immunoblot demonstrates expression of ATF4. (C) Mapping of arginine residues found in ATF4. (D) ATF4 is methylated by PRMT1 at amino acid (aa) residue 239. GST-purified ATF4 WT and various ATF4 deletion mutants (top panel) and arginine mutants (bottom panel) were subjected to in vitro methylation assays together with purified PRMT1 and S-adenosyl methionine as a methyl donor. Proteins were resolved by SDS-PAGE, stained with Coomassie blue (ATF4 input), dried and analyzed by autofluorography. Mutation of arginine 239 abolishes methylation of ATF4.

Figure 4. Loss of PRMT1-mediated methylation at residue R239 attenuates ATF4 function

(A) Both WT and ATF4-R239K are equally expressed. Non-methylated ATF4-R239K was generated by overlap extension PCR (see Materials and Methods). WT and ATF4-R239K as well as an empty vector (EV) were retrovirally transduced into immortalized Atf4^−/− MEFs. Protein lysates were generated and subjected to IP with ATF4Ab. Immunoblot demonstrates expression of ATF4. β-actin levels were used as a loading control. (B) ATF4 methylation potentiates its activity. EV, ATF4-WT and ATF4-R239K complemented MEFs were subjected to qPCR to determine expression of ATF4 targets identified by gene expression analysis (Table 3). Data are presented as fold induction of mRNA (expression level of untreated EV MEFs were set to 1). Bars represent average data from four independent experiments ± SEM. P-values are indicated with ***P < 0.001, **P < 0.01 and *P < 0.05 (two-tailed paired t-test). (C) WT MEFs were treated with 50 μM AMI-1 in the presence and absence of glutamine and western blot was performed for several ATF4 targets. GAPDH levels were used as a loading control.

Figure 5. The absence of Btg1 expression in mouse bone marrow-derived B-cell progenitors enhances cell survival in ASNase-treated cells

(A) Btg1 expression is upregulated upon ASNase treatment. WT CD19-positive bone marrow cells were treated with 0.5 U/ml ASNase for 48 h. Btg1 mRNA levels were measured by qPCR and the relative expression compared to untreated cells (set to 1) is shown. Bars represent average data from four independent experiments ± SEM. (B) WT and Btg1^−/− CD19-positive bone marrow cells were treated with 0.5 U/ml ASNase for 48 h. The metabolic activity of the cells was measured by MTS assay and the relative cell survival compared to untreated cells (set as 100%) is shown. Bars represent average data from six independent experiments ± SEM. p-values are indicated with ***P < 0.001 (two-tailed t-test).

Figure 6. Model of BTG1-mediated ATF4 regulation during cellular stress

(A) In response to sustained stress conditions, BTG1 recruits PRMT1 to methylate ATF4, - indicated by ‘Me’- which promotes transcription of a subset of ATF4 target genes, leading to increased apoptosis. (B) In the absence of BTG1, PRMT1 no longer binds to and methylates ATF4, shifting the balance from pro-apoptotic to pro-survival. As a consequence, loss of BTG1 function promotes (tumor) cell survival.