Fig 2: Modular pleiotropy within protein complexes resolves C7orf26 as a member of the Integrator complex INTS10–13-14 module(A) STAGA and ATAC protein complexes share a histone acetyltransferase module. From cancer cell fitness data alone, Webster inferred separate functional effects of STAGA and ATAC depletion while decomposing the effect of knocking out shared subunits as a mixture of STAGA and ATAC depletion.(B) The SWI/SNF family protein complexes consist of specialized subunits bound to the common enzymatic subunit SMARCA4. From cancer cell fitness data alone, Webster inferred separate functional effects of depleting each subcomplex while decomposing the effect of SMARCA4 knockout as affecting all three subcomplexes.(C) The Integrator complex is a modular RNA endonuclease complex. From cancer cell fitness data alone, Webster learned three distinct functional effects involving Integrator complex componentry. Each of the three functions captured the specific effect of knocking out recently discovered structural modules of the complex. An unknown interactor, C7orf26, was approximated as a mixture of all three Integrator functions, with the strongest loaded function mapping to the INTS10–13-14 functional module.(D) Schematic of two C7orf26 gene splice variants curated by the genotype-tissue expression (GTEx) portal.(E) Immunoprecipitation (IP) of C7orf26-HA variants from 293T nuclear extracts, immunoblotted for INTS10–13-14 module subunits.(F) Mass spectrometry of C7orf26-HA and C7orf26-V5 immunoprecipitations from 293T cells shows stoichiometric pull down of INTS10, 13, and 14 with the full-length C7orf26.(G) 293T cells with CRISPR-Cas9 perturbation of C7orf26 display loss of INTS10 at the protein level.(H) IP of endogenous C7orf26 from 293T cells with and without INTS10 knockout.(I) Model figure—C7orf26 stabilizes INTS10, which assembles together with the INTS13–14 heterodimer to form the INTS10–13-14-C7orf26 module.

Fig 3: INTS10 forms a module with INTS13–INTS14 in cells and binds to the MIDAS pocket of the INTS14 VWA domain.a Size exclusion chromatography of nuclear extract from HEK293T cells. All even fractions were analyzed by Western blotting. INTS1 and INTS7 served as marker proteins for the INT core, while INTS4–9–11 form the cleavage module. Elution points of molecular weight markers are indicated on the top. In addition, elution volumes of purified INTS13–INTS14 and INTS10–INTS13–INTS14 are indicated. b, c Coprecipitation of HA-INTS10 from HEK293T cells with V5-SBP-INTS13 b or V5-SBP-INTS14 c. Wild-type and binding patch double mutants were compared. Coprecipitation of endogenous INTS14 or INTS13 is detected as control for complex disruption. Inputs (αV5-blots 1%, αHA-blots 0.38% and 0.75%, αINTS14-blot 0.38%, αINTS13-blot 0.03%) and bound fractions (αV5-blots 3%, αHA-blots 20 and 15%, αINTS14/αINTS13-blots 20%) were analyzed by Western blotting. d Deletion constructs of V5-SBP-INTS14 were tested for copurification of HA-INTS10 and endogenous INTS13. Inputs (αV5-blot 1%, αHA-blot 1%, αINTS13-blot 0.1%) and bound fractions (αV5-blot 4%, αHA-blot 5%, αINTS13-blot 20%) were characterized by Western blotting. e Copurification of MBP-INTS10 with GST-INTS14-VWA from E. coli lysates. GST served as negative control. f Structure superposition (r.m.s.d. 3.6 Å over 160 residues) of the VWA domain of INTS14 (green) with the VWA domain of integrin-αL (ITGAL, gray, PDB-ID: 1T0P73) in complex with intercellular adhesion molecule 3 (ICAM3, purple). The black square highlights the position of the MIDAS pocket, residues lining the pocket are shown as sticks. g Focused view of the MIDAS pockets of both VWA domains highlights conservation of the site in INTS14. The Mg2+-ion in the ITGAL pocket that coordinates a Glu of the interaction partner ICAM3 is shown as a gray sphere. Ion coordination is indicated by black dotted lines. Coloring as in f. Residues mutated in INTS14 are underlined. h Coprecipitation of HA-INTS10 from HEK293T cells with V5-SBP-INTS14 wt or MIDAS pocket mutants (D8A, S10A, S12A [DA2SA] and L11E, R15A [LERA]). V5-SBP-MBP served as control. Inputs (αV5-blots 1%, αHA-blots 0.75%) and bound fractions (αV5-blots 3%, αHA-blots 15%) were analyzed by Western blotting.

Fig 4: INTS13 and INTS14 form an interlinked heterodimer.a Coprecipitation experiments of λN-HA-tagged INTS14 with V5-SBP-tagged INTS13 from HEK293T cells using streptavidin beads. V5-SBP-MBP served as a negative control. Inputs (αV5-blot 1%, αHA-blot 0.38%) and bound fractions (αV5-blot 7.5%, αHA-blot 20%) were analyzed by Western blotting. b Same experiment as in a but with λN-HA-INTS13 and V5-SBP-INTS14. c Copurification of INTS14 with 2S-tagged INTS13 from insect cells expressing both proteins. The identity of proteins on the Coomassie-stained gel was verified by Western blotting. The asterisk marks a C-terminal degradation product of INTS13. C-terminal truncations of INTS13 are not detected by the specific antibody. d Domain organization of INTS13 and INTS14 as observed in the crystal structure of the complex. The dotted line marks portions of both proteins that are visible in the structure. CMBM marks the cleavage module-binding motif identified in this study. e Structure of the INTS13–INTS14 complex colored as in d. The two-fold rotation axis (C2) for the pseudosymmetry is indicated by a black oval. The position of the chain interlink is marked by a black box. The cartoon provides a simplified schematic for the domain arrangement of the complex. f Close-up of the interlink between INTS13 (dark orange) and INTS14 (light blue). g DSS crosslinks between Lys residues of INTS13 and INTS14 as determined by XL–MS. Crosslinks between residues that are visible in the crystal structure are shown, including crosslinks within the INTS13–INTS14 complex (blue) and crosslinks stemming most likely from complex oligomerization (red). Domain schemes are colored as in d and every 100th residue is marked with a tick. L indicates linker regions.

Fig 5: INTS10–13–14 are required for termination after pausing.a Scheme of the RNAPII pause-termination reporter used in b. Reduced RNAPII pausing and termination at the HIV-1 TAR element results in elongation and increased RLuc levels. b Luciferase assay using the TAR pause-termination reporter. Cells depleted of individual INT subunits were transfected with the TAR reporter and FLuc transfection control plasmid. Statistical significance was determined for each knockdown condition relative to control. Bars show mean values of biological triplicates together with the individual replicates as dots. Error bars represent standard deviations. Statistical significance was determined by unpaired, two-tailed t-test: not significant (ns) p ≥ 0.05; significant *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Western blots in Fig. 6a. Source data are provided as a Source Data file. c Summary model of INT modules and their respective protein interaction networks. INTS10–INTS13–INTS14 brings cleavage module and target RNA into proximity.