The USP11/TCEAL1 complex promotes transcription elongation to sustain oncogenic gene expression in neuroblastoma

Abstract

During early transcription, RNA polymerase II (RNAPII) undergoes a series of structural transitions controlled by cyclin-dependent kinases. How protein ubiquitylation and proteasomal degradation control the function of RNAPII is less well understood. Here we show that the deubiquitinating enzyme USP11 forms a complex with TCEAL1, a member of the TFIIS (TCEA)-like protein family. TCEAL1 shares sequence homology with the RNAPII interaction domain of the elongation factor TFIIS (which controls the fate of backtracked RNAPII) and competes with TFIIS for binding to core promoters. USP11 protects TCEAL1 from proteasomal degradation, and TCEAL1 recruits USP11 to RNAPII. Both USP11 and TCEAL1 promote transcription elongation and maintain expression of RPB8, an essential subunit of all three nuclear RNA polymerases. In neuroblastoma, USP11- and TCEAL1-dependent genes define a gene expression program that is characteristic for mesenchymal tumors, which are described as able to escape from many treatments, suggesting that the USP11/TCEAL1 complex promotes transcription elongation to support a critical oncogenic gene expression program.

Keywords: MYCN; RNA polymerase II; TCEAL1; TFIIS; USP11; transcription elongation.

Figure 1.

USP11 interacts with multiple proteins involved in basal transcription. (A) Volcano plot of the consensus USP11 interactome with selected proteins labeled in blue. The X-axis displays the log₂ ratio of proteins in HA-USP11 immunoprecipitations (IPs) relative to control IPs. The Y-axis shows the −log₁₀ of P-value for each protein. The vertical dashed line indicates a log₂ ratio of 2, and the horizontal dashed line indicates a P-value of 0.05 (n = 4; unless specified otherwise, n indicates the number of independent biological replicates). (B) Scatter plot of mass spectrometry data analysis of anti-HA IPs from SH-EP neuroblastoma cells stably expressing N-terminally (NT) or C-terminally (CT) HA-tagged USP11. The graph shows the normalized log₂ ratio over anti-HA IP performed in vector control cells (ctr). Dashed lines indicate a log₂ ratio of 2 (n = 2 for each IP). Proteins labeled in blue were validated in the IPs shown below. (C) Immunoblots of anti-USP11 IPs from SH-EP cells. The input corresponds to 1.5% of the amount used for the precipitation. Nonspecific IgG and GAPDH were used as controls (n = 2). (D) Immunoblots of anti-USP7 IPs from SH-EP cells. The input corresponds to 1.5% of the amount used for the precipitation. Nonspecific IgG and GAPDH show controls (n = 2). (E) Immunoblots of anti-TCEAL1 IPs from SH-EP cells. The input corresponds to 2% of the amount used for the precipitation. Nonspecific IgG and GAPDH show controls (n = 2). (F) Immunoblots of anti-USP11 and anti-USP7 IPs in the indicated cell lines. The input corresponds to 1.5% of the amount used for the precipitation. Nonspecific IgG and GAPDH show controls (n = 2). (G) Immunoblots of anti-USP7 (top) and anti-USP11 (bottom) IPs in SH-EP cells expressing the indicated doxycycline-inducible shRNAs in the presence or absence of 1 µg/mL doxycycline for 48 h. The input corresponds to 1.5% of the amount used for the precipitation. Nonspecific IgG and GAPDH show controls (n = 2). (H) Immunoblots of the indicated proteins in SH-EP cells upon knockdown of USP11 using inducible shRNA by adding 1 µg/mL doxycycline for 48 h. Cells were treated with 10 µM MG132 for 4 h where indicated. VCL was used as a loading control (n = 3). (I) Quantification of the TCEAL1 level of biological replicates as described in H from two different shRNAs. Data are shown as mean ± SD. P-values were calculated using a paired two-sided t-test (n = 5).

Figure 2.

USP11 promotes transcription elongation. (A) Immunoblots of anti-USP11 (left) and anti-USP7 (right) IPs in SH-EP cells expressing the indicated doxycycline-inducible shRNAs in the presence or absence of 1 µg/mL doxycycline for 48 h. The input corresponds to 1.5% of the amount used for the precipitation. Nonspecific IgG and GAPDH show controls (n = 2). (B) Genome browser tracks of the LDHA locus showing chromatin association of RNAPII in SH-EP cells upon shRNA-mediated knockdown of USP11 or siRNA-mediated knockdown of USP7. (C) Global average read density of RNAPII ChIP-Rx at the TSS ± 2 kb, treated as described in B. Data show mean ± SEM (N = genes analyzed, n = 2). (D) Traveling ratio of RNAPII. Ratios were calculated for cells described in B for either USP11 (left) or USP7 (right) depletion. P-value was calculated using a two-sided Kolmogorov–Smirnov test (N = 8071 for USP11, N = 10,787 for USP7). (E) Genome browser tracks of the LDHA locus showing chromatin association of pS2-RNAPII in SH-EP cells as described in B. (F) Global average read density of pS2-RNAPII ChIP-Rx at the TES ± 2 kb, treated as described in B. Data show mean ± SEM (N = genes analyzed, n = 2). (G) Immunoblots of anti-USP11 IPs from SH-EP MYCN-ER cells expressing doxycycline-inducible shRNA targeting TCEAL1 in the presence or absence of 1 µg/mL doxycycline for 48 h (“−TCEAL1”) and MYCN activation with 200 nM 4-OHT for 4 h (“+MYCN”). The input corresponds to 1.5% of the amount used for the precipitation. Nonspecific IgG and GAPDH were used as controls (n = 3).

Figure 3.

Chromatin association of TCEAL1 depends on the TFIIS homology region. (A) Genome browser tracks of the indicated loci showing chromatin association of HA-tagged TCEAL1 and RNAPII in SH-EP cells. Anti-HA chromatin immunoprecipitations (ChIPs) from cells that do not express HA-tagged TCEAL1 were used as control (ctr) (n = 2). (B) Global average read density of TCEAL1 and RNAPII ChIP-Rx at the TSS ± 2 kb. Data show mean ± SEM (N = genes analyzed). (C) Heat maps indicating global chromatin occupancy of TCEAL1 and RNAPII at active genes at the TSS ± 5 kb. The heat map is sorted according to RNAPII binding. (TSS) Transcription start site, (R) biological replicate. N = 16,285 genes analyzed. (D) Venn diagram showing overlap of TCEAL1 and RNAPII occupancy at the TSS ± 500 bp. (E) Sequence alignment of TFIIS and TCEAL1, with the domains of the TFIIS protein shown at the top and those of TCEAL1 shown at the bottom. The turquoise color in the HTH indicates the homology domain with TFIIS (THS). (RS) Arginine/serine-rich domain, (ZnF-L) zinc finger-like domain, (HTH) helix–turn–helix domain. Amino acids subjected to mutagenesis are highlighted in red. (F) AlphaFold2 prediction of TCEAL1 structure. Predicted local distance difference test (pLDDT) indicates modeling confidence and is marked by different shades of blue. Zoomed-in view shows the α helix conserved between TCEAL1 and TFIIS. Amino acids subjected to mutagenesis are labeled. (G) Immunoblots of SH-EP cells expressing HA-tagged TCEAL1 wild type or the indicated mutant alleles. VCL was used as a loading control (n = 2). (H) Immunoblots of TCEAL1 IPs of SH-EP cells expressing HA-tagged TCEAL1^wt, TCEAL1^3RA, or TCEAL1^5RA. Coprecipitated proteins are indicated. Nonspecific IgG and GAPDH were used as controls (n = 3). (I) Anti-HA ChIP at the indicated promoter regions in SH-EP cells expressing HA-tagged TCEAL1 wild type or the indicated mutant alleles. IgG was used as a negative control. Shown is the mean of technical replicates of one representative experiment (n = 3).

Figure 4.

TCEAL1 and TFIIS compete for binding to RNAPII. (A) TFIIS ChIP at the indicated loci in SH-EP cells expressing a doxycycline-inducible shRNA targeting TCEAL1 (1 µg/mL doxycycline for 48 h). IgG was used as negative control. Shown is the mean of technical triplicates of one representative experiment (n = 3). (B) Genome browser tracks of the indicated loci showing TFIIS ChIP-Rx data in cells as described in A (n = 2). (C) Global average read density of TFIIS ChIP-Rx at the TSS ± 2 kb. Data show mean ± SEM (N = genes analyzed). (D) Box plot representing global reads of TFIIS at the TSSs upon TCEAL1 knockdown. The gray line indicates the mean of unperturbed cells. P-value was calculated using a paired two-sided Student's t-test with unequal variance (N = 21,005 genes analyzed). (E) Density plot showing the change in ratio of TFIIS to RNAPII occupancy at the TSS ± 2 kb upon TCEAL1 knockdown (N = genes analyzed). (F) Representative PLA images showing detected PLA spots (top) and PLA signal merged with Hoechst 33342 staining (bottom). Scale bar, 20 µm. n = 3. (G) Box plot of single-cell analysis of nuclear proximity ligation assay (PLA) foci between TFIIS and differently phosphorylated RNAPII of one representative PLA experiment. P-values were calculated using an unpaired t-test (n = 3). (H) Immunoblots of RNAPII IPs of SH-EP cells expressing HA-tagged TCEAL1^wt or TCEAL1^3RA. Coprecipitated proteins are indicated. Nonspecific IgG and VCL were used as control (n = 2). (I) Global average read density of TFIIS ChIP-seq data at the TSS ± 2 kb in cells upon stable expression of TCEAL1^wt or TCEAL1^3RA. Data show mean ± SEM (N = genes analyzed, n = 2).

Figure 5.

USP11 maintains RPB8 expression. (A) Volcano plot of mass spectrometry data of IMR-5 cells expressing a doxycycline-inducible shRNA targeting USP11 (1 µg/mL doxycycline for 48 h). The X-axis displays log₂ FC of proteins in cells with control levels (ctr) or knockdown of USP11 (−USP11). The Y-axis shows the −log₁₀ P-value for each protein (n = 2). (B, left) Immunoblots from cells described in A. VCL was used as a loading control (n = 4). (Right) Quantification of USP11 and RPB8 levels of biological replicates. Data are shown as mean ± SD. P-values were calculated using a paired two-sided t-test. (C) Volcano plot showing changes of different RNAP subunits after knockdown of USP11. The axis is as described in A. (D) Table showing log₂ FC of different RNAPII subunits after shRNA-mediated depletion of USP11. (E, left) Representative images of SH-EP cells immunostained for RPB8 or RPB1 after knockdown of USP11. Scale bar, 50 µm. (Right) Box plots of single-cell analysis of a representative experiment. P-values were calculated using an unpaired t-test (n = 3). (F, left) Representative images of SH-EP cells immunostained for RPB8 after knockdown of TCEAL1. Scale bar, 50 µm. (Right) Box plots of single-cell analysis of a representative experiment. P-values were calculated using an unpaired t-test (n = 3).

Figure 6.

TCEAL1 promotes RNAPII association and elongation. (A) Global average read density plot at TSS ± 2 kb of HA ChIP-seq data in SH-EP cells expressing HA-tagged TCEAL1 after treatment with 200 nM THZ1 for 8 h, 1 µM NVP-2 for 3 h, or DMSO as control. SH-EP cells were used as negative control (ctr). Data show mean ± SEM (N = genes analyzed, n = 2). (B) Immunoblots of the indicated proteins in SH-EP MYCN-ER cells expressing doxycycline-inducible shRNA targeting TCEAL1 (1 µg/mL doxycycline for 48 h; “−TCEAL1”) and MYCN activation (200 nM 4-OHT for 4 h; “+MYCN”). VCL was used as a loading control (n = 3). (C) Genome browser tracks of the FASN locus showing chromatin association of RNAPII in SH-EP MYCN-ER cells (treatment as in B). (D) Global average read density of RNAPII ChIP-Rx data at TSS ± 2 kb as described in B. Data show mean ± SEM (N = genes analyzed, n = 2). (E) Genome browser tracks of the FASN locus showing chromatin association of pS2-RNAPII in SH-EP MYCN-ER cells as described in B (n = 2). (F) Pausing index of RNAPII. Ratios were calculated for cells described in B either without (left) or with (right) MYCN induction. P-value was calculated using a two-sided Kolmogorov–Smirnov test (N = 22,725 genes analyzed).

Figure 7.

USP11 and TCEAL1 promote a mesenchymal gene expression program. (A) Correlation of gene sets upon depletion of TCEAL1 and USP11. Each dot represents one gene set. Dark gray indicates FDR < 0.25 in at least one analysis. Gene sets reflecting TGF-β signatures are colored in red. (B) Correlation of the expression of TCEAL1-dependent (left) or USP11-dependent (right) target genes with mesenchymal gene expression program in n = 498 primary neuroblastoma samples (N = 100 genes analyzed). (C) Heat map comparing adrenergic (blue) and mesenchymal (orange) cells versus the most significant downregulated genes upon depletion of TCEAL1 (for gene names, see Supplemental Fig. S7B). R-values comparing expression of genes of each cell line with the Foroutan TGF-β EMT up signature are depicted at the bottom. (D) Correlation of the expression of genes in SH-EP (left) or IMR-5 (right) upon siRNA-mediated knockdown of TCEAL1 with mesenchymal gene expression program in n = 498 primary neuroblastoma samples (FDR 0.05, N = 3159 for SH-EP, N = 132 for IMR-5). (E) Correlation of the expression of genes in SH-EP (left) or IMR-5 (right) upon siRNA-mediated knockdown of TCEAL1 with adrenergic gene expression program in n = 498 primary neuroblastoma samples (FDR 0.05, N = 3286 for SH-EP, N = 135 for IMR-5). (F) Speculative model of our findings: We propose that the USP11/TCEAL1 complex promotes RNAPII elongation by buffering TFIIS activity and maintaining RPB8 levels, potentially by counteracting its BRCA1-dependent ubiquitylation.