Transcriptional coactivator PGC-1α contains a novel CBP80-binding motif that orchestrates efficient target gene expression

Abstract

Although peroxisome proliferator-activated receptor-γ (PPARγ) coactivator 1α (PGC-1α) is a well-established transcriptional coactivator for the metabolic adaptation of mammalian cells to diverse physiological stresses, the molecular mechanism by which it functions is incompletely understood. Here we used in vitro binding assays, X-ray crystallography, and immunoprecipitations of mouse myoblast cell lysates to define a previously unknown cap-binding protein 80 (CBP80)-binding motif (CBM) in the C terminus of PGC-1α. We show that the CBM, which consists of a nine-amino-acid α helix, is critical for the association of PGC-1α with CBP80 at the 5' cap of target transcripts. Results from RNA sequencing demonstrate that the PGC-1α CBM promotes RNA synthesis from promyogenic genes. Our findings reveal a new conduit between DNA-associated and RNA-associated proteins that functions in a cap-binding protein surveillance mechanism, without which efficient differentiation of myoblasts to myotubes fails to occur.

Keywords: CBP80; PGC-1α; cap-binding complex; gene transcription; myogenesis; pre-mRNP quality control.

Figure 1.

Defining the α-helical CBM of PGC-1β. (A) Diagram of human PGC-1β denoting some of the known or putative functional regions. (RD) Negative regulatory domain. Lines below the diagram show Escherichia coli-produced GST-PGC-1β regions tested for the ability to pull down baculovirus-produced CBP80. Numbers specify amino acids. (B) SDS–polyacrylamide separation and Coomassie blue staining of GST pull-downs of E. coli-produced GST alone (−) as a negative control or the denoted GST-tagged regions of human PGC-1β in the presence of baculovirus-produced human CBP80. The asterisk denotes the fragment of PGC-1β that binds CBP80. (C) Crystal structure of PGC-1β–CBC–m⁷GpppA, where the box encompasses the PGC-1β–CBP80 interface that is enlarged at the left. (Green) CBP80; (blue) CBP20; (black) m⁷GpppA; (golden brown) PGC-1β peptide, which includes the nine (FDSLLKEAQ) amino acids at residues 1011–1019 that comprise the CBM. The residues of CBP80 and PGC-1β that interact are detailed and labeled. CBM residue A1018 was omitted because it is hidden under the α-helix ribbon. (D) Diagrams (as in A) of human PGC-1β wild-type [PGC-1β(WT)] and PGC-1β(ΔCBM). (E) Western blots of lysates of MCF-7 cells transiently transfected with plasmid (p) encoding the specified Flag-tagged PGC-1β variant or Flag alone (−) before or after anti-Flag (α-Flag) immunoprecipitation, the latter in the presence (+) or absence (−) of RNase I. Here and in the following Western blots, calnexin served to control for variations in loading and immunoprecipitation specificity, and the top left triangle denotes threefold serial dilutions of samples to use for quantitative comparisons. (F) Diagrams of human CBP80 wild type, CBP80(Δ1–44), and CBP80(9Mut) denoting some of the known or putative functional regions (Shatsky et al. 2014). (MIF4G) Middle domain of eukaryotic initiation factor 4G (eIF4G). (G) Western blots of lysates of MCF-7 cells transiently transfected with plasmid encoding the specified Flag-tagged CBP80 variant or Flag alone (−) before or after anti-Flag immunoprecipitation, the latter in the presence (+) or absence (−) of RNase I. Here and in the following Western blots, β-Actin served to control for variations in loading and immunoprecipitation specificity.

Figure 2.

PGC-1α interacts with the CBC largely via its CBM. (A) Diagram of human PGC-1α and its regions tested for the ability to pull down CBP80, as in Figure 1A. (LXXLL motif) α-Helical sequence that binds NRs, where XX represent two lysines in PGC-1α. (B) GST pull-down analyses of CBP80 with GST alone or GST-tagged regions of human PGC-1α, as in Figure 1B. (C) Western blots of total cell lysates or nuclear lysates of C2C12 myoblasts (MBs) before or after immunoprecipitation, the latter in the presence (+) or absence (−) of RNase I, using anti-PGC-1α or, as a control, rabbit IgG (rIgG). (D) Western blots of lysates of C2C12 MBs transiently transfected with plasmid producing Flag-CBP80(WT) (+) or Flag alone (−) either before immunoprecipitation, after a first anti-Flag immunoprecipitation, or after a second immunoprecipitation using anti-PGC-1α or, as a control, rIgG. Samples from the second immunoprecipitation were loaded so that the amounts of PGC-1α in the first and second immunoprecipitations were equivalent. (E) Diagrams (as in A) of human PGC-1α(WT) and variants. (F,G) Western blots of lysates of C2C12 MBs transiently transfected with plasmid encoding the specified Flag-tagged PGC-1α variant or Flag alone (−) before or after anti-Flag immunoprecipitation, the latter in the presence (+) or absence (−) of RNase I. (KD) Knockdown.

Figure 3.

PGC-1α interacts with the cap of newly synthesized target transcripts largely via its CBM. (A,B) Histogram representation of RT-qPCR quantitations of pre-mRNA (A) or mRNA (B) from three PGC-1α-responsive genes (Idh3b, Pfk1, and Sirt5) and, as negative controls, two unresponsive genes (Hprt and β-Actin) after immunoprecipitation relative to before immunoprecipitation using anti-PGC-1α (red) or, as a control, rIgG (blue), where values after rIgG immunoprecipitation relative to before immunoprecipitation are set to 1. (C,D) As in A and B but using lysates of C2C12 cells treated as in Figure 2D and after a first anti-Flag immunoprecipitation or a second immunoprecipitation using anti-PGC-1α or, as a control, rIgG, where values in the first anti-Flag immunoprecipitation in the presence of Flag relative to before immunoprecipitation and values after the second immunoprecipitation using rIgG relative to before the second immunoprecipitation are set to 1. (E,F) As in A and B but after anti-Flag immunoprecipitation using lysates of PGC-1α knockdown C2C12 MBs transiently transfected with plasmid encoding the specified Flag-PGC-1α variant or Flag alone (−), where values after anti-Flag-PGC-1α(WT) immunoprecipitation relative to before immunoprecipitation were set to 1. (G) Schematic for antisense DNA oligonucleotide-directed RNase H-mediated cleavage of cellular pre-mRNP and mRNP. The mixture of three oligonucleotides, each complementary to nucleotides 6–22 of the denoted transcript, directs RNase H-mediated cleavage so as to remove the 5′ cap structure (CAP), including the associated proteins, from the body of the transcript. Assays for RNase H cleavage are presented (Supplemental Fig. S3I). (H,I) As in E and F but after anti-Flag immunoprecipitation, the latter in the presence of the oligonucleotide mix and either RNase H (+) or no RNase H (−), where values after anti-Flag-PGC-1α(WT) immunoprecipitation in the absence of RNase H relative to before immunoprecipitation were set to 1. For all histograms, results are means ± SD. n ≥ 3; n = 2 for H and I. (*) P < 0.05; (**) P < 0.01; (no asterisks) P > 0.05 (i.e., not significant) compared with control immunoprecipitation (A–D), compared with PGC-1α knockdown C2C12 MBs expressing Flag-PGC-1α(WT) (E,F), or compared with minus RNase H treatment (H,I) by a two-tailed unpaired Student's t-test.

Figure 4.

The CBM controls the transcriptional activity of PGC-1α. (A,B) Histogram representation of RT-qPCR quantitations of pre-mRNA (A) or mRNA (B) from the three PGC-1α-responsive genes normalized to the level of Hprt mRNA using lysates of PGC-1α knockdown C2C12 MBs treated as in Figure 2, F and G, or Supplemental Figure S2D. (C,D) As in A and B but using lysates of C2C12 MBs transiently cotransfected with either control or CBP80 siRNA and with plasmid encoding the specified siRNA-resistant Flag-CBP80 variant or Flag alone (−). pre-mRNA and mRNA levels were normalized to the level of U1 snRNA. (E) Heat map of hierarchically clustered mRNA expression from RNA sequencing (RNA-seq) analyses of the specified cells: wild-type C2C12 MBs expressing either Flag (−) or PGC-1α knockdown C2C12 MBs expressing Flag (−), Flag-PGC-1α(WT), or Flag-PGC-1α(ΔCBM). Shown are genes whose expression was significantly affected by PGC-1α knockdown and significantly rescued by Flag-PGC-1α(WT) and/or Flag-PGC-1α(ΔCBM). The color key represents row-scaled expression values. For all histograms, results are means ± SD. n = 3. (*) P < 0.05; (**) P < 0.01 by a two-tailed unpaired Student's t-test.

Figure 5.

PGC-1α CBM promotes myogenesis by inducing myogenic gene transcription. (A) GO analysis of C2C12-MB genes whose up-regulation is dependent on the PGC-1α CBM (Supplemental Table S3). Asterisks specify myogenesis-related GO terms. (B,C) Histogram representation of RT-qPCR quantitations of PGC-1α CBM-dependent myogenic pre-mRNAs (B) or mRNAs (C) identified in Supplemental Figure S5A in cells treated as in Figure 4, A and B. (D,E) Histogram representation of RT-qPCR quantitations of PGC-1α CBM-dependent promyogenic pre-mRNAs (D) or mRNAs (E) using lysates of PGC-1α knockdown MBs expressing Flag-PGC-1α(WT) or Flag-PGC-1α(ΔCBM) before (0 d) and after (1, 3, and 5 d) culturing in differentiation medium (DM). (F) Western blots showing the expression of exogenous Flag-PGC-1α(WT) or Flag-PGC-1α(ΔCBM) and endogenous PGC-1α, CBP80, and myogenic markers in C2C12 MBs treated as in D and E. For all RT-qPCR analyses, results are means ± SD. n = 3 for B and C; n = 2 for D and E. (*) P < 0.05; (**) P < 0.01; (no asterisks) P > 0.05 between the indicated conditions (B,C) or comparing PCG-1α knockdown C2C12 MBs expressing Flag-PGC-1α(ΔCBM) with Flag-PGC-1α(WT) (D,E) by a two-tailed unpaired Student's t-test.