The splicing factor U2AF small subunit is functionally conserved between fission yeast and humans

Abstract

The small subunit of U2AF, which functions in 3' splice site recognition, is more highly conserved than its heterodimeric partner yet is less thoroughly investigated. Remarkably, we find that the small subunit of Schizosaccharomyces pombe U2AF (U2AF(SM)) can be replaced in vivo by its human counterpart, demonstrating that the conservation extends to function. Precursor mRNAs accumulate in S. pombe following U2AF(SM) depletion in a time frame consistent with a role in splicing. A comprehensive mutational analysis reveals that all three conserved domains are required for viability. Notably, however, a tryptophan in the pseudo-RNA recognition motif implicated in a key contact with the large subunit by crystallographic data is dispensable whereas amino acids implicated in RNA recognition are critical. Mutagenesis of the two zinc-binding domains demonstrates that they are neither equivalent nor redundant. Finally, two- and three-hybrid analyses indicate that mutations with effects on large-subunit interactions are rare whereas virtually all alleles tested diminished RNA binding by the heterodimer. In addition to demonstrating extraordinary conservation of U2AF small-subunit function, these results provide new insights into the roles of individual domains and residues.

FIG. 1.

Analysis of a diploid S. pombe strain heterozygous for the uaf2::LEU2 deletion-replacement allele. (A) Genomic Southern blot of DNA from SpCW1 (see Materials and Methods) and its parent strain, SpDS3. The probe was pBSK⁺, containing the 5.1-kb XbaI fragment labeled by random priming. In SpCW1, replacement of one copy of the uaf2⁺ gene with LEU2 increases the size of the endogenous XbaI fragment (5.1 kb) by 1.4 kb, producing a 6.5-kb fragment containing the uaf2::LEU2 null allele. The second allele of uaf2⁺ remains a 5.1-kb fragment. In the parental diploid SpDS3 (3), both copies migrate at 5.1 kb. (B) Schematic representations of the wild-type and deletion-replacement alleles of uaf2⁺. (C) Plate assay demonstrating complementation of the fission yeast small-subunit null allele by plasmids expressing either S. pombe or human U2AF³⁵. Empty vector was used as a negative control. A, adenine; L, leucine. See the legend to Table 2 for details.

FIG. 2.

Effects of metabolically depleting either subunit of S. pombe U2AF. (A) Growth and splicing data for Uaf2p (U2AF^SM). (Top panel) Growth curves in the presence (+B₁) or absence (−B₁) of thiamine for a haploid strain in which the uaf2⁺ gene disruption is complemented by a plasmid carrying wild-type uaf2⁺ expressed from the thiamine-repressible nmt1 promoter. Abs₆₀₀, absorbance at 600 nm. (Middle panel) RT-PCR splicing assays on cdc16-I2 in cells harboring uaf2⁺ under thiamine control. Total RNA was isolated from the depletion strain at the time points indicated (in hours) after addition of B₁ and at the first time point in the absence of B₁. Total RNA was also extracted from the control haploid strain SpDS2 (4) at an A₆₀₀ of 1.0. Mock RT-PCRs lacking reverse transcriptase indicated no contamination with genomic DNA (data not shown). (Bottom panel) Quantitation of the splicing assays depicted in the middle panel expressed as percentages of precursor accumulation [precursor RNA/(precursor mRNA + mature mRNA) × 100]. The levels of precursor and mature message were determined using Bio-Rad Quantity One gel documentation software (version 4.1.1). Error bars indicate standard deviations for three splicing assays. (B) Results are presented as described for panel A except that the data are for a strain in which Uaf1p (U2AF^LG) was undergoing metabolic depletion. (C) Sequence of the cdc16-I2 pre-mRNA from the branchpoint (italic characters) to three bases downstream of the 3′ AG (uppercase characters). Designations for pyrimidines between the branchpoint and 3′ splice site are underlined.

FIG. 3.

Sequence alignment of S. pombe U2AF^SM and human U2AF³⁵, indicating the locations of mutations analyzed in this study. The alignment was produced using the ClustalW program with minor manual refinement. *, amino acid identity; •, amino acid similarity. The putative zinc-coordinating residues are highlighted in white characters on a black background for each ZBD. The ΨRRM is highlighted in black characters on a gray background with the RNP2-like (amino acids 46 to 51) and RNP1-like (amino acids 104 to 111) peptides boxed. On the basis of the results of our structure-function studies (see text) together with those obtained in a structure-based alignment of several RRM-RNA complexes (49), we shifted the position of the RNP2-like peptide by one amino acid relative to the results shown in Fig. 2 of reference . The single and multiple amino acid substitutions analyzed here (with their phenotypic consequences denoted according to the key) are indicated above the S. pombe sequence.

FIG. 4.

Two-hybrid analyses of protein-protein interactions between wild-type U2AF^LG and wild-type and mutant U2AF^SM. The figure shows the results of β-galactosidase assays, with the data expressed as percentages of the wild-type U2AF^SM-U2AF^LG interactions. For reference, the phenotypic consequences for S. pombe of these alleles are indicated according to the key for Fig. 3. Error bars indicate standard deviations (n = 3).

FIG. 5.

(A) Three-dimensional ribbon diagram of the domain of the human U2AF minimal heterodimer. (B) Close-up image of the U2AF³⁵ ΨRRM region proposed to contact RNA. The β-sheet containing the RNP1-like peptide is colored yellow, and the β-sheet containing the RNP2-like peptide is shown in red. Amino acids corresponding to those mutated in the present study are shown as stick figures to allow the orientations of side chains to be visualized. Numbers and identities of the amino acids in the small subunit of S. pombe U2AF are given in parentheses where they differ from the human ortholog. The images were prepared using RasMol software (version 2.7.2.1) (42) and the coordinates supplied by Kielkopf et al. (24).

FIG. 6.

Plate assays to test RNA binding in the modified RNA three-hybrid system. The strength of the RNA-protein interaction is reflected by growth in the presence of increasing amounts of 3-AT (9, 44). (Top panel) Effects of mutations in or near the RNP2-like motif of the U2AF^SM ΨRRM. (Bottom panel) Effects of mutations in putative zinc ligands of U2AF^SM. Phenotypes for S. pombe are indicated as described for Fig. 3. The sequence of the test RNA (I-1D4), which corresponds to the 3′ end of an S. pombe intron (from just downstream of the branchpoint to 10 bases beyond the 3′ AG), is shown at the bottom. All assays were repeated with essentially identical results. L, leucine; U, uracil; H, histidine.

FIG. 7.

Two-hybrid analyses of protein-protein interactions between small-subunit orthologs and SR proteins from S. pombe and humans. The figure shows the results of β-galactosidase assays (expressed in Miller units on a logarithmic histogram). The protein combination assayed in each case is indicated beneath the histogram. The y axis is adjusted to the background activity observed in the host strain harboring empty bait and prey vectors. Error bars indicate standard deviations (n = 3).