Integrator subunit 4 is a 'Symplekin-like' scaffold that associates with INTS9/11 to form the Integrator cleavage module

Abstract

Integrator (INT) is a transcriptional regulatory complex associated with RNA polymerase II that is required for the 3'-end processing of both UsnRNAs and enhancer RNAs. Integrator subunits 9 (INTS9) and INTS11 constitute the catalytic core of INT and are paralogues of the cleavage and polyadenylation specificity factors CPSF100 and CPSF73. While CPSF73/100 are known to associate with a third protein called Symplekin, there is no paralog of Symplekin within INT raising the question of how INTS9/11 associate with the other INT subunits. Here, we have identified that INTS4 is a specific and conserved interaction partner of INTS9/11 that does not interact with either subunit individually. Although INTS4 has no significant homology with Symplekin, it possesses N-terminal HEAT repeats similar to Symplekin but also contains a β-sheet rich C-terminal region, both of which are important to bind INTS9/11. We assess three functions of INT including UsnRNA 3'-end processing, maintenance of Cajal body structural integrity, and formation of histone locus bodies to conclude that INTS4/9/11 are the most critical of the INT subunits for UsnRNA biogenesis. Altogether, these results indicate that INTS4/9/11 compose a heterotrimeric complex that likely represents the Integrator 'cleavage module' responsible for its endonucleolytic activity.

Figure 1.

Multiple modified yeast two-hybrid screens identify INTS4 as an INTS9/11 interacting protein. (A) (Upper panel) Human INTS9 is fused to the Gal4 DNA binding domain (DB) and is co-expressed with human INTS11 that is not fused to any yeast protein. This combination of INT subunits was then screened against an activation domain library containing each of the mammalian INT subunits. (Lower panel) Results of the modified yeast two-hybrid screen identifying INTS4 as the only INT subunit capable of supporting growth on selection media. AD stands for fusion to Gal4 activation domain and ‘ϕ’ symbol means AD with no INT subunit fused to it. The images on the left column are a serial dilution series (5-fold increments) of yeast that were plated onto permissive media lacking leucine, tryptophan, and uracil (-L/-W/-U). The images on the right column are the same as the left except the media also lacked histidine and was supplemented with 5 mM 3-amino-1,2,4-triazole (-L/-W/-U/-H +3AT). (B) (Upper panel) Similar to panel A except human INTS11 is fused to the Gal4 DNA binding domain and co-expressed with human INTS9 that is not fused to any yeast protein and was screened against an activation domain library containing each of the mammalian INT subunits. (C and D) Same as panels A and B except both screens were conducted using INT subunits derived from Drosophila melanogaster.

Figure 2.

INTS4 likely recognizes the INTS9/11 heterodimeric interface to mediate interaction. (A) Results of modified yeast two-hybrid where one of the subunits is eliminated at each position. ‘BD’ represents the INT subunit fused to the DNA binding domain; ‘trans’ represents the INT subunit expressed as a non-fused protein; ‘AD’ represents the INT subunit expressed as an activation domain fusion protein. The images on the left column are a serial dilution series (5-fold increments) of yeast that were plated onto permissive media lacking leucine, tryptophan, and uracil (-L/-W/-U). The images on the right column are the same as the left except the media also lacked histidine and was supplemented with 5 mM 3-amino-1,2,4-triazole (-L/-W/-U/-H +3AT). (B) Same as panel A except using Drosophila INT cDNA components. (C) Results of modified yeast two-hybrid using INTS9 as the DNA binding fusion protein but with the introduction of a mutation (mt) that disrupts the INTS9/11 heterodimeric interface (arginine-644 to a glutamate). The mutation within INTS9 was determined through a previous crystal structure of INTS9/11 and validated experimentally to abrogate INTS9 binding to INTS11 (34). (D) Same as in panel C except the position of INTS11 and INTS9 are switched and the mutation is introduced into INTS11. The mutation within INTS11 (L509A/F511A) was previously determined to disrupt the INTS9/11 heterodimeric interface (34). (E) Results of co-immunoprecipitation of INTS4/9/11 from cell lysates derived from transfected 293T cells. In all three panels, cells were transfected with either HA-INTS4 or empty HA vector. In the left panel, cells were also transfected with both FLAG-INTS9 and myc-INTS11 that were wild-type. Input represents 5% of the input cell lysate and pulldowns using anti-HA affinity resin were from either empty HA transfected cells or HA-INTS4 cells. In the pulldowns, 40% of the material was loaded into the gel. The middle and right panels are the same as the left panel except in each case, either INTS9 or INTS11 wild type cDNA was exchanged with a mutant cDNA used in panels C or D. (F) Results of size exclusion chromatography of purified INTS4-INTS9-INTS11 heterotrimer. The recombinant proteins were produced using a multigene baculovirus where only INTS4 possessed an N-terminal His tag. Following purification using nickel beads, the heterotrimeric complex was loaded onto a gel filtration column and the eluted fractions were monitored for protein content using A280. The peak fractions were analyzed by SDS-PAGE and visualized by Coomassie blue staining (gel inset).

Figure 3.

Regions within the N- and C-terminus are required for INTS4 binding and are also essential for INTS4 function in UsnRNA processing. (A) Schematic of human Symplekin and human INTS4 highlighting known and predicted domains. (B) Results of modified yeast two-hybrid assay where INTS4 HEAT repeats are deleted at two HEAT repeat intervals. The terms: ‘BD’ stands for the Gal4 DNA binding domain; ‘trans’ refers to the protein expressed in yeast that is not fused to any other protein; ‘AD’ refers to the Gal4 activation domain. The left panel are images of yeast grown on non-selective media plates (-L/-W/-U) while the right panel are images of yeast grown on selective media plates (-L/-W/-U/-H +3AT). (C) Same as panel B except regions at the C-terminus of INTS4 are being deleted at 100 amino acid increments. (D) Same as panel C except smaller deletions from the C-terminus of INTS4 are tested. (E) Same as panel D except mutations where five amino acids at a time are being changed to alanine in the context of the full-length protein. The abbreviation of ‘A/959–963’ refer to the amino acids within INTS4 protein that were changed to alanine. (F) Western blot analysis of cell lysates transfected with the U7-GFP reporter after being treated with either control siRNA (Con.) or two distinct INTS4 siRNA (4–1 or 4–2). In addition to reporter and siRNA, cells were also either transfected with empty vector (VA) or with RNAi-resistant INTS4 cDNA (+wt*). Results shown are representative of triplicate biological experiments. (G) Western blot analysis of cell lysates treated with control siRNA or INTS4 siRNA that were also transfected with the U7-GFP reporter and either empty vector or INTS4 cDNA that is RNAi-resistant. The RNAi-resistant INTS4 cDNA used were either wild-type or housing deletions in the HEAT repeats as labeled. Results shown are representative of triplicate biological experiments. (H) Similar to panel G except alanine mutants of INTS4 were tested as described in panel E. Results shown are representative of triplicate biological experiments.

Figure 4.

Depletion of INT subunits 4, 9, and 11 lead to the greatest amount of UsnRNA misprocessing. (A) Western blot analysis of HeLa cells transfected with either two different siRNA that each target eight individual INT subunits or a control siRNA (Con.). In each blot, probing with an antibody raised against tubulin is used as a loading control. (B) Upper panel- fluorescence microscopy image is representative of observations upon knockdown of INTS4. HeLa cells were treated with either control siRNA or with siRNA targeting INTS4 and then cotransfected with a vector encoding mCherry and the U7-GFP reporter. The U7-GFP reporter only expresses if there is misprocessing of the U7snRNA gene in response to loss of INT function. Lower panel is a Western blot analysis of cell lysates from cells transfected with the U7-GFP reporter and the mCherry expression plasmid. In each case, cells were first transfected with siRNA targeting individual INT subunits or a control siRNA. The bottom panel is a loading control where tubulin was probed. Lower panel is the results of fluorescence plate reader assays of triplicate transfections with either siRNA targeting INT subunits. Results are representative of triplicate experiments.

Figure 5.

Localization of endogenous INT subunits. (A) Western blot analysis of nuclear extract derived from HeLa cells. In each case, equal amounts of nuclear extract was loaded and antibodies were identical to those used in Figure 4A. The lower band observed for INTS1 is a degradation product and not a nonspecific band. (B) Confocal microscopy of HeLa cells fixed and immunostained with antibodies used in panel A and anti-coilin antibodies. DAPI (blue) is included in INTS1, 4, 9 and 11 to distinguish between nuclear and cytoplasmic INT subunit localization. Arrows indicate CBs surrounded by nearby INT subunit-positive foci, likely representing UsnRNA gene loci associated with CBs. Cytoplasmic signal observed in several of the INT localizations is likely due to nonspecific signal from the antiserum as supported by cell fractionation experiments shown in Supplementary Figure S3.

Figure 6.

INTS4/9/11 are required for Cajal Body integrity. (A) Confocal microscopy of HeLa cells fixed and immunostained with antibodies raised against either coilin or Gemin2, the member of the SMN complex, that have been transfected with siRNA targeting INT subunits or control siRNA (Con. or SLBP). Nuclei are visualized using DAPI staining (blue). Arrows indicate CBs in cells unaffected by specific INT subunit depletion. (B) Similar to panel A except cells were only treated with either control siRNA or siRNA targeting INTS9 and stained for coilin, SMN, or fibrillarin. (C) Quantification of 100 cells visualized from panel A demonstrating the percentage of cells with positive nucleolar coilin staining. (D) Western blot analysis of cell lysates from cells treated with either control siRNA or INTS11 siRNA.

Figure 7.

INTS4/9/11 are required to maintain histone locus bodies. (A) Confocal microscopy of HeLa cells fixed and immunostained with antibodies raised against either coilin or NPAT, a specific histone transcription factor, that have been transfected with siRNA targeting INT subunits or control siRNA (Con. or SLBP). Nuclei are visualized using DAPI staining (purple). (B) Quantification of 100 cells visualized from panel A demonstrating the percentage of cells lacking HLBs detected by NPAT staining. (C) Similar to panel A except HeLa cells were only treated with either control siRNA or siRNA targeting INTS4 and stained for coilin and FLASH, a specific histone pre-mRNA processing factor, indicating that HLBs were truly disassembled in INTS4 kd cells.

Figure 8.

Model of the INTS4/9/11 Cleavage Module in UsnRNA 3′-end processing. Integrator associates with the CTD tail of RNA polymerase II to promote cotranscriptional processing of pre-UsnRNA immediately adjacent to the Cajal body. The INTS4 subunit utilizes both N- and C-terminal domains to interact with the INTS9/11 heterodimer to cause the cleavage of pre-UsnRNA. The heterotrimer of INTS4/9/11 associates with the rest of INT through interaction with a yet-to-be identified INT subunit.