1 Table S1 Cryo-EM data collection, refinement and validation statistics Data collection and processing CPSF-160 WDR33 (EMDB-7114) (PDB 6BM0) CPSF-160 WDR33 (EMDB-7113) (PDB 6BLY) CPSF-160 WDR33 CPSF-30 PAS RNA (EMDB-7112) (PDB 6BLL) Magnification 22,500 22,500 22,500 Voltage (kv) 300 300 300 Electron exposure (e /Å 2 ) 47 70 70 Defocus range (μm) 1.3~2.7 1.2~2.5 1.2~2.5 Pixel size (Å) 1.30 1.07 1.07 Symmetry imposed C1 C1 C1 Image stacks (no.) 1625 2486 2486 Initial particle images (no.) 372,707 1,144,122 1,144,122 Final particle images (no.) 205,373 456,310 173,632 Map resolution (Å) FSC threshold 3.78 0.143 3.36 0.143 3.42 0.143 Map sharpening B-factor (Å 2 ) 180 186 157 Refinement Number of protein residues 1,538 1,540 1,669 Number of RNA nucleotides Number of metal ions 0 0 0 0 7 3 Number of atoms 12,244 12,263 13,435 R.m.s. deviations Bond lengths (Å) Bond angles ( ) 0.008 1.05 0.008 0.98 0.010 1.02 PDB validation Clash score Poor rotamers (%) Ramachandran plot Favored (%) Allowed (%) Disallowed (%) 18 0.52 91.06 7.81 1.13 7 0.22 91.20 8.47 0.33 10 0.28 90.48 9.33 0.18
Fig. S1. Image processing of the CPSF-160 WDR33 complex. (A). Area of an image of the complex in negative stain. Scale bar: 50 nm. (B). Selected 2D class averages obtained with the ISAC algorithm. Side length of individual averages: 19 nm. (C). The initial model generated with the VIPER algorithm using the ISAC averages obtained for the negatively stained complex. (D). Area of an image of the vitrified complex. Some particles are circled. Scale bar: 50 nm. (E). Selected 2D class averages obtained with RELION-2. Side length of individual averages: 21 nm. (F). Image-processing workflow for 3D classification and refinement in RELION-2 that resulted in a density map at 3.8 Å resolution. See Methods for details. 2
Fig. S2. Image processing of the CPSF-160 WDR33 CPSF-30 PAS RNA complex. (A). Area of a cryo-em image of the vitrified complex. Some particles are circled. Scale bar: 50 nm. (B). Selected 2D class averages obtained with RELION-2. Side length of individual averages: 21 nm. (C). Image-processing workflow for 3D classification and refinement in RELION-2 that resulted in a density map at 3.42 Å resolution. See Methods for details. 3
Fig. S3. FSC curves and local densities. (A). Gold-standard FSC curves calculated between independently refined half maps for (1) the density map of the CPSF-160 WDR33 complex (labeled 160-33, yellow), (2) the density map obtained with the CPSF-160 WDR33 CPSF-30 PAS RNA sample by combining five classes after the first 3D classification (Map 1 in Supplementary Fig. 2) (red), and (3) the density map obtained with the CPSF-160 WDR33 CPSF-30 PAS RNA sample by combining five classes after the second 3D classification (Map 2 in Supplementary Fig. 2) (green). (B). Cross-validation FSC curves: green, refined model versus half map 1 used for refinement (work map); yellow, refined model versus half map 2 not used for refinement (free map); red, refined model versus the combined final map. The similarity of the work and free curves suggests no substantial over-fitting. (C). Representative cryo-em densities for Map2 obtained with the CPSF-160 WDR33 CPSF-30 PAS RNA sample. (D). Two views of the cryo-em map of the quaternary complex, segmented and colored as in panel a of Fig. 1. The views are related by 90 rotation around the vertical axis. (E). Two views of the cryo-em map of the binary complex at 3.8 Å resolution. 4
5 Fig. S4. Sequence alignment of selected WDR33 homologs. The secondary structure elements in the human WDR33 structure are shown. Residues in contact with CPSF-160, CPSF-30, and PAS RNA are indicated with purple, green, and orange dots, respectively. The boundaries of the WD40 domain are indicated. Only residues 1-451 of human WDR33 are included in the alignment. Hs: Homo sapiens, Mm: Mus musculus, Xt: Xenopus tropicalis; Dr: Danio rerio, Dm: Drosophila melanogaster.
Fig. S5. Overlay of the structures of the quaternary and binary complexes. (A). Overlay of the structures of the CPSF- 160 WDR33 CPSF-30 PAS RNA quaternary complex (in color) and the CPSF-160 WDR33 binary complex (gray). (B). Close-up view of the N-terminal segment of WDR33 containing residues 43-54 (cyan). This segment is disordered in the binary complex (gray). 6
Fig. S6. Sequence alignment of selected CPSF-30 homologs. The secondary structure elements in the human CPSF- 30 structure are shown. Residues in contact with CPSF-160, WDR33, and PAS RNA are indicated with purple, light blue, and orange dots, respectively. The boundaries of ZF1-ZF3 are indicated. Hs: Homo sapiens, Mm: Mus musculus, Xt: Xenopus tropicalis; Dr: Danio rerio, Dm: Drosophila melanogaster. 7
Fig. S7. Comparison of the RNA-binding mode in ZF2 of CPSF-30 with that in other CCCH zinc fingers. (A). Overlay of the structures of CPSF-30 ZF2 and Nab2 ZF6 (PDB ID 5L2L). (B). Overlay of the structures of CPSF- 30 ZF2 and MBNL1 ZF3 (PDB ID 3D2S). (C). Overlay of the structures of CPSF-30 ZF2 and TIS11d ZF1 (PDB ID 1RGO). 8
Fig. S8. Gel filtration profiles for purified mpsf proteins used for biochemical studies. Gel filtration profiles for the CPSF-160 WDR33 CPSF-30 ternary complex (red), the CPSF-160 WDR33 binary complex (blue), CPSF-160 alone (cyan), MBP-CPSF-30 fusion protein (magenta), and a mixture of CPSF-160 and MBP-CPSF-30 (orange) are shown. Inset: SDS-PAGE gel of the protein samples. The second peak for the ternary complex contains MBP and excess CPSF-30 (lane 5). 9
Fig. S9. Structural similarity between the CPSF-160 WDR33 complex and the DDB1 DDB2 complex. (A). Overlay of the structures of the CPSF-160 WDR33 CPSF-30 PAS RNA quaternary complex (in color) with that of the DDB1 DDB2 dsdna complex (gray, PDB ID 3EI1). The overlay is between CPSF-160 and DDB1 only. (B). Overlay of the structures of the CPSF-160 WDR33 CPSF-30 PAS RNA quaternary complex (in color) with that of the DDB1 CUL4A RBX1 SV5-V complex (gray, PDB ID 2HYE). The overlay is based on BPA and BPC of CPSF-160. The protein bound between BPA and BPC of DDB1 (SV5-V) does not contain a b-propeller. The bottom face of BPB mediates interaction with CUL4A. 10