Activation of the Endonuclease that Defines mRNA 3' Ends Requires Incorporation into an 8-Subunit Core Cleavage and Polyadenylation Factor Complex

Abstract

Cleavage and polyadenylation factor (CPF/CPSF) is a multi-protein complex essential for formation of eukaryotic mRNA 3' ends. CPF cleaves pre-mRNAs at a specific site and adds a poly(A) tail. The cleavage reaction defines the 3' end of the mature mRNA, and thus the activity of the endonuclease is highly regulated. Here, we show that reconstitution of specific pre-mRNA cleavage with recombinant yeast proteins requires incorporation of the Ysh1 endonuclease into an eight-subunit "CPF_core" complex. Cleavage also requires the accessory cleavage factors IA and IB, which bind substrate pre-mRNAs and CPF, likely facilitating assembly of an active complex. Using X-ray crystallography, electron microscopy, and mass spectrometry, we determine the structure of Ysh1 bound to Mpe1 and the arrangement of subunits within CPF_core. Together, our data suggest that the active mRNA 3' end processing machinery is a dynamic assembly that is licensed to cleave only when all protein factors come together at the polyadenylation site.

Keywords: X-ray crystallography; baculovirus; cleavage; cryo-EM; hydrogen-deuterium exchange; mRNA; mass spectrometry; nuclease; polyadenylation; pre-mRNA.

Graphical abstract

Figure 1

The Mpe1 UBL Domain Binds to the Ysh1 Catalytic Domain Next to the Active Site Tunnel (A) Domain diagram of Ysh1 and Mpe1 proteins, with truncations indicated by black lines. ZnK, zinc knuckle. (B) SDS-PAGE analysis of pull-down experiments following baculovirus-driven co-expression of pairs of Ysh1 and Mpe1 constructs shown in (A), with full-length Cft2. The tagged Mpe1 constructs (asterisks) were captured by Strep-Tactin resin, and co-purification of Ysh1 and Cft2 was analyzed. (C) SDS-PAGE analysis of complexes identified in (B) after anion exchange chromatography. The Ysh1-Mpe1 proteins remain associated, but Cft2 dissociates. (D) X-ray crystal structure of the Ysh1 N-terminal catalytic domain (yellow) bound to the Mpe1 N-terminal UBL domain (orange). N and C termini of both models are indicated, zinc-coordinating residues are shown in sticks, and zinc ions are spheres. A slice through the complex (right) reveals a narrow tunnel leading to a large solvent-filled cavity adjacent to the active site. Inset: electrostatic surface potential at pH 7.4. A large basic patch comprising residues from both proteins lies adjacent to the active site tunnel. (E) Details of metal ion coordination in the Ysh1 active site. (F) Details of the interface between Ysh1 and Mpe1. Hydrogen bonds and electrostatic interactions are indicated by green dashed lines. Two orthogonal views are shown. See also Figure S1.

Figure 2

Mass Spectrometry and Cryo-EM Define the Interactions among Full-Length Ysh1, Mpe1, and Yjr141w (A) Interactions among Ysh1, Mpe1, and Yjr141w mapped by crosslinking mass spectrometry of the Ysh1-Mpe1-Yjr141w trimer and Ysh1-Mpe1 and Ysh1-Yjr141w heterodimers. Lines are color-coded as indicated. (B) Hydrogen-deuterium exchange mass spectrometry difference plot (Ysh1-Mpe1-Yjr141w versus Ysh1-Yjr141w) showing peptides of Ysh1 that are protected (negative) and exposed (positive) by Mpe1. (C) Hydrogen-deuterium exchange mass-spectrometry difference plot (Ysh1-Mpe1-Yjr141w versus Ysh1-Mpe1) showing peptides of Ysh1 that are protected (negative) and exposed (positive) by Yjr141w. In (B) and (C), triplicate data from four independent color-coded time-points are shown. The significance threshold is indicated by a dotted line. Gray shading indicates the SD of all charge states and replicates per peptide. (D) Cryo-EM analysis of the Ysh1-Mpe1-Yjr141w heterotrimer. A representative micrograph at original magnification × 105,000 and −0.5 μm defocus. (E) Selected 2D class averages of aligned particles. (F) The crystal structure from Figure 1D was docked into the EM map filtered to 6 Å resolution. No density was observed for the Ysh1 CTD or the Yjr141w or Mpe1 CTDs. See also Figure S2 and Table S1.

Figure 3

Ysh1 Is Primed for Activation by Assembly into an Eight-Subunit CPF_core Complex (A) Schematic diagrams showing the expression and purification workflow, composition of recombinant complexes, and details of the in vitro activity assay. Proteins are represented by circles, with a yellow star to highlight an enzymatic subunit. S, StrepII tag; H, His₆ tag; CPF_pol, polymerase module. (B) SDS-PAGE analysis of recombinant protein complexes after affinity, anion exchange, and size exclusion chromatography. Asterisks indicate contaminant proteins. (C) The CYC1 model pre-mRNA is specifically cleaved by CPF_core with CF IA and CF IB, and the 5ʹ-cleavage product is polyadenylated in the presence of ATP, as shown by denaturing gel electrophoresis of RNA. The negative control reaction (−) contained CF IA and CF IB, but not CPF_core. (D) Denaturing RNA gel electrophoresis of cleavage assay time courses performed using the protein complexes shown in (B). The negative control lanes (−) show no RNA cleavage when incubated with CF IA and CF IB (left) or buffer (right) for 90 min. See also Figure S3.

Figure 4

CPF_core Binds and Cleaves a 36-nt Minimal RNA Substrate (A) Sequences of RNA substrates derived from the CYC1 3ʹ UTR. Each substrate carries both 5ʹ-FAM and 3ʹ-A647 labels (red and blue stars, respectively). The canonical cleavage site is highlighted in bold, and the minimal sequence required for efficient cleavage is represented by the gray box. (B) Denaturing gel electrophoresis of the short RNA substrates after incubation with CPF_core, CF IA, and CF IB. The negative control reaction (−) contained CF IA and CF IB, but not CPF_core. (C) Electrophoretic mobility shift assays (EMSAs) performed with CYC1d (cleaved by CPF_core) and CYC1f (not cleaved by CPF_core) RNAs. See also Figures S4–S6 and Table S2.

Figure 5

The Enzymatic Subunits of CPF_core Assemble around a Central Scaffold (A) Size-exclusion chromatography of CPF_core and SDS-PAGE analysis of fractions across the peak. Asterisks indicate contaminant proteins. (B) Representative negative-stain micrograph of CPF_core. (C) Negative-stain 2D class averages show a distinctive 21-nm particle with the polymerase module at one end. (D) Representative cryo-EM micrograph of CPF_core. (E) Selected 2D class averages from cryo-EM analysis of CPF_core. Approximately 80% of the particles are present in classes that comprise the 13-nm scaffold of the polymerase module only. Up to three additional subunits are visible in ∼0.5% of the particles. (F) A model for the structure of CPF_core obtained from a 3D reconstruction of the negative-stain data. Three orthogonal views filtered to 25 Å are shown as insets. The cryo-EM structure of Cft1-Pfs2-Yth1 (Casañal et al., 2017) and X-ray crystal structures of Cft2 (Mandel et al., 2006), Pap1-Fip1 (Meinke et al., 2008), and Ysh1-Mpe1 (this work; Figure 1) are docked into the negative-stain map. Known disordered or flexible regions are indicated with colored lines. The weak interaction between Ysh1 and Cft2 CTDs is indicated with dashed lines. Also see Figure S7.

Figure 6

Model for 3ʹ End Formation on the Minimal CYC1 Pre-mRNA Substrate (A) CF IA, CF IB, and CPF each preferentially bind certain RNA sequences. CF IA binds U-rich elements via interactions with Rna15 RRM domains, CF IB binds UA-rich sequences, and CPF_core binds the 5ʹ AAGAA element. (B) When all of the correct sequence elements are present, the 3ʹ end processing machinery can assemble into an active complex, resulting in an opening of the active site cleft of Ysh1. Pre-mRNA cleavage occurs within a 3-nt window. Some of the interactions in the model are speculative.