Functional analysis of Hsh155/SF3b1 interactions with the U2 snRNA/branch site duplex

Abstract

SF3b1 is an essential component of the U2 snRNP implicated in branch site (BS) recognition and found to be frequently mutated in several human cancers. While recent structures of yeast and human SF3b1 have revealed its molecular architecture, the importance of specific RNA:protein contacts and conformational changes remains largely uncharacterized. Here, we performed mutational analysis of yeast SF3b1, guided by recent structures of the spliceosome. We find that conserved amino acids contacting the U2 snRNA backbone of the U2/BS duplex are nonessential, and that yeast can tolerate truncation of the HEAT repeats containing these amino acids. The pocket housing the branchpoint adenosine (BP-A) is also amenable to mutation despite strong conservation. However, mutations that support viability can still lead to defects in splicing pre-mRNAs with nonconsensus BS substitutions found at -3, -2, -1, and +1 positions relative to the BP-A or at the branchpoint position. Through the generation of yeast and human chimeric proteins, we further defined the functionally conserved regions of Hsh155 as well as identify changes in BS usage resulting from inclusion of human SF3b1 HEAT repeats. Moreover, these chimeric proteins confer a sensitivity to small molecule inhibition by pladienolide B to yeast splicing. Together, these data reveal the importance of individual contacts of Hsh155/SF3b1 to the U2/BS duplex and define their contribution to BS usage by the spliceosome.

Keywords: SF3b1; pladienolide B; snRNP; spliceosome; splicing.

FIGURE 1.

Alanine scan of Hsh155 contacts to the U2/BS duplex. (A) Cartoon representation of the HEAT domains of Hsh155 (gray) bound to the U2/BS duplex (blue/red) from a cryo-EM structure of an activated spliceosome (pdb 5GM6) (Yan et al. 2016). The region of human SF3b1 frequently mutated in MDS is colored in green. (B) Schematic illustration of Hsh155-U2/BS duplex contacts observed in the structure shown in A. (C) Sequence alignment of the U2/BS duplex binding site of Hsh155 from yeast (Sc) and human (Hs). (D) Growth of haploid yeast containing the indicated Hsh155 mutations on nonselective (−Trp) and selective (−Trp +FOA) media at 30°C.

FIGURE 2.

Identification of the essential regions of S. cerevisiae Hsh155. (A) Schematic illustration of Hsh155 domains. Numbers and arrows indicate the positions of truncation mutants shown in panels B and C. (B,C) Growth of haploid yeast containing the indicated Hsh155 N-terminal (panel B) and C-terminal (panel C) truncations on nonselective (−Trp) and selective (−Trp +FOA) media at 30°C. (D) Cartoon illustration of the HEAT domains of Hsh155 (gray) bound to the U2/BS duplex (blue/red) from a cryo-EM structure of an activated spliceosome (pdb 5GM6). The nonessential region is shown in yellow.

FIGURE 3.

Mutation of BP-A pocket alters nonconsensus BS usage. (A) Schematic illustration of the ACT1-CUP1 reporter pre-mRNA and assay. The consensus 5′SS, 3′SS, and BS are shown in the intron. The position of A258 is noted and the branchpoint (BP-A; A259) is underlined. (B) Results from a Cu²⁺ growth assay of strains containing mutations to U2/BS duplex contacts and either the consensus ACT1-CUP1 reporter plasmid or one with a nonconsensus A258U BS substitution. (C) Results from a Cu²⁺ growth assay of strains carrying mutations to the BP-A pocket and containing a nonconsensus A258U BS reporter plasmid. (D) Heatmap summarizing ACT1-CUP1 reporter data for the indicated Hsh155 mutants and BS reporters. Plotted data represent the log₂ transform of the ratio of the maximum [Cu²⁺] at which growth was observed for the indicated mutant to the maximum [Cu²⁺] at which growth was observed for WT Hsh155. Purple indicates decreased growth relative to WT, while orange indicates improved growth. In panels B and C, each bar represents the average of three independent experiments, and error bars represent the standard deviation.

FIGURE 4.

Humanized Hsh155 functions in yeast and alters nonconsensus BS usage. (A) In vitro splicing assay with WT Hsh155 in the presence or absence of PB (20 µM). PB does not inhibit yeast splicing at this concentration. (B) Schematic illustration of the humanized Hsh155 chimeric proteins. (C,D) Growth of haploid yeast containing the indicated humanized Hsh155 proteins on nonselective (−Trp) and selective (−Trp +FOA) media at 30°C. (E) Heatmap summarizing ACT1-CUP1 reporter data for the indicated humanized Hsh155 proteins and MDS mutants with several BS reporters. Plotted data represent the log₂ transform of the ratio of the maximum [Cu²⁺] at which growth was observed for the indicated mutant to the maximum [Cu²⁺] at which growth was observed for WT Hsh155. Purple indicates decreased growth relative to WT, and orange indicates improved growth. (F) Heatmap of ACT1-CUP1 reporter data as in panel E but including the Hs5-16 humanized protein with and without the R1074H splicing inhibitor resistance mutation (note that numbering refers to human SF3b1).

FIGURE 5.

PB inhibits yeast splicing with humanized Hsh155. (A) In vitro splicing assay with WT Hsh155 or the indicated humanized Hsh155 variants in the presence or absence of PB (500 nM). (B) Quantification of in vitro splicing data represented by panel A. Each bar represents the average of three independent experiments, and error bars represent the standard deviation. Data were normalized to the extent of splicing observed for each individual strain in the absence of PB. (C) Titration of PB inhibition of in vitro splicing of extracts containing Hs5-16. (D) Quantification of the data shown in panel C. Approximately 50% inhibition (∼IC₅₀) is observed at 25 nM PB.

FIGURE 6.

U2/BS duplex recognition by SF3b1 and PHF5A during splicing. A properly closed conformation of SF3b1/Hsh155 is associated with accommodation of the bulged, BP-A of the U2/BS duplex within the hydrophobic pocket of SF3b1/Hsh155 (dark gray) and alignment of the phosphodiester backbone of the intron strand of the U2/BS duplex with a set of positively charged amino acids within the C-terminal HEAT domain. N-terminal HEAT repeats and PHF5A/Rds3 contact the U2/BS duplex. Formation of this complex promotes selection of a particular BS and splicing (pbd: 5GM6) (Yan et al. 2016). Note that Rds3 was omitted from the inset for clarity.