Systematic analysis of the protein interaction network for the human transcription machinery reveals the identity of the 7SK capping enzyme

Abstract

We have performed a survey of soluble human protein complexes containing components of the transcription and RNA processing machineries using protein affinity purification coupled to mass spectrometry. Thirty-two tagged polypeptides yielded a network of 805 high-confidence interactions. Remarkably, the network is significantly enriched in proteins that regulate the formation of protein complexes, including a number of previously uncharacterized proteins for which we have inferred functions. The RNA polymerase II (RNAP II)-associated proteins (RPAPs) are physically and functionally associated with RNAP II, forming an interface between the enzyme and chaperone/scaffolding proteins. BCDIN3 is the 7SK snRNA methylphosphate capping enzyme (MePCE) present in an snRNP complex containing both RNA processing and transcription factors, including the elongation factor P-TEFb. Our results define a high-density protein interaction network for the mammalian transcription machinery and uncover multiple regulatory factors that target the transcription machinery.

Figure 1. A Method for the Analysis of Protein Complexes in Human Cells

(A) Flow chart of the procedure used for the affinity purification and mass spectrometry identification of protein complexes in HEK293 cells. (B) Computational validation of protein interactions. Sensitivity/specificity ROC curve, parameterized by the IR score, for the discrimination between literature-supported interactions and dubious interactions (see Table S1). The chosen IR score threshold results in a sensitivity of 83% and a specificity of 83%. The IR was calculated based on the method described in the Supplemental Experimental Procedures. (C) Summary of the overall screen for protein interactions. (D) List of the 32 TAP-tagged polypeptides used in this study.

Figure 2. A Network of Protein Complexes Involving the RNAP II Basal Transcription Machinery

Overview of the interaction map showing the 805 validated interactions obtained using 32 tagged polypeptides. The interactions are represented as directional edges extending from the tagged protein using the web-based software VisANT (http://visant.bu.edu). The thickness of the line that connects two proteins is proportional to the IR score obtained for this interaction (see Table S1). The nodes are colored according to their GO annotations (http://www.geneontology.org). The tagged proteins are indicated.

Figure 3. Previously Uncharacterized Protein Complexes Containing RNAP II in Association with Regulators of Protein Complex Assembly

(A) Network highlighting the interactions of RPAPs-XAB1 with RNAP II, the regulatory complexes Integrator and Mediator, and a group of proteins with chaperone/scaffolding activity. (B) Elution profile of RNAP II and the RPAPs-XAB1 in gel filtration. (C) Previously uncharacterized proteins are associated with RNAP II. For each human protein, a representation with putative functional domains provided by the NCBI’s conserved domain database, the name of the closest yeast homologous gene as determined by a BLASTP search on the June 2005 assembly version of the UCSC genome browser, and gene essentiality in yeast provided by the SGD are shown. ARM, ARM repeat protein superfamily; ATP_bind_1, conserved hypothetical ATP-binding protein; DUF408, domain of unknown function; TPR, tetratricopeptide repeat domain; WD40, WD40 domain. (D) Clustering analysis of mRNA expression profiles with tet-promoter alleles corresponding to RBA50, NPA3, YLR243W, and RPB11, in comparison to profiles from 214 other tet-promoter mutants representing diverse cellular functions (Mnaimneh et al., 2004).

Figure 4. A Complex Comprising the Elongation Factor P-TEFb, the 7SK and U6 snRNAs, and the Previously Uncharacterized Protein BCDIN3

(A) Network highlighting the interactions of BCDIN3 and CDK9 with various RNA processing factors, P-TEFb (cyclin T1/cyclin T2 subunits), and its regulatory factors HEXIM1 and HEXIM2. (B) Gel filtration analysis showing that P-TEFb, HEXIM1, and the RNA processing factor SART3 cofractionate with BCDIN3. (C) The 7SK and U6 snRNAs are present in the BCDIN3-TAP eluate. RNA blots were performed on total RNA extracted from HeLa cells (T; 900 ng) or RNA extracted from the BCDIN3-TAP eluate (B; 60 ng) and probed with RNA oligos specific for U2, U6, or 7SK snRNAs. Migration of the three intact RNA species is indicated. (D) The 7SK snRNP complex is partly resistant to extensive digestion with RNase A. SDS gel showing BCDIN3 affinity eluates prepared from cell extracts treated or not with RNase A. Western blots confirming the presence of BCDIN3, P-TEFb, HEXIM1, and SART3 in both eluates are shown.

Figure 5. BCDIN3 Is a Conserved Methyltransferase that Targets the 7SK snRNA

(A) Linear representation of BCDIN3 with its RNA methylase and Bin3 domains provided by the NCBI’s conserved domain database. The amino acid sequence of the C-terminal region of BCDIN3 containing the putative S-adenosyl methionine (AdoMet)-binding domain is shown. The signature seven β strand methyltransferase motifs I, Ia, II, and III are indicated. (B) Evolutionary conservation of human BCDIN3. A schematic representation of BCDIN3 and its protein homolog according to the NCBI’s Homolo-Gene system is provided. The conserved regions containing the methyltransferase motifs I, Ia, II, and III are indicated. (C) The 7SK snRNA is methylated in vitro by BCDIN3. Recombinant 7SK snRNA was incubated with His-tagged BCDIN3 and ³H-AdoMet in the presence of the buffer alone, AdoHcy, or RNase A. The ³H-methylated RNA product was detected by autoradiography. (D) Immunodepletion of BCDIN3 impairs the 7SK methylation activity of whole-cell extracts and addition of recombinant BCDIN3-His to the depleted extract rescues 7SK methylation. HEK293 whole-cell extracts (WCE) were incubated or not with protein A Sepharose beads coupled to either pre-immune or anti-BCDIN3 serum, and supernatants were immunoblotted with antibodies directed against BCDIN3 and SART3 (as a loading control). WCE incubated or not with beads coated with preimmune or anti-BCDIN3 serum were assayed for methylation of recombinant 7SK snRNA in the absence or presence of His-tagged BCDIN3. RNA was visualized by ethidium bromide (EB) staining (lower panel), and the ³H-methylated product was detected by autoradiography (upper panel).

Figure 6. BCDIN3 Is the 7SK snRNA MePCE

(A) Dephosphorylation of the 7SK snRNA impairs its methylation by BCDIN3. Recombinant 7SK snRNA was treated with (+) or without (−) calf intestinal alkaline phosphatase (CIAP) prior to its incubation with His-tagged BCDIN3 and ³H-AdoMet. The RNA was visualized by ethidium bromide (EB) staining (lower panel), and the methylated product was detected by autoradiography (upper panel). A schematic representation of the assay is shown. (B) Enzymatic decapping of BCDIN3-methylated 7SK removes the radiolabeled methyl group. Recombinant 7SK was incubated with His-tagged BCDIN3 and ³H-AdoMet and then decapped or not (−) with different amounts of tobacco acid (TA) pyrophosphatase. The RNA was visualized by ethidium bromide (EB) staining (lower panel), and the methylated product was detected by autoradiography (upper panel). A schematic representation of the assay is shown.

Figure 7. BCDIN3 Silencing Decreases the Steady-State Level of Cellular 7SK

BCDIN3 silencing was monitored by western blotting using extracts from HEK293 cells mock transfected or transfected with control or BCDIN3 siRNA (tubulin was used as a loading control). 7SK steady-state levels were assessed by RNA blotting using total RNA extracted from HEK293 cells mock transfected or transfected with control or BCDIN3 siRNA (48 hr and 72 hr posttransfection) and probed with RNA oligonucleotides that specifically detect 7SK, U6, or U2 snRNAs.