Deep mutational scanning of an RRM domain of the Saccharomyces cerevisiae poly(A)-binding protein

Abstract

The RNA recognition motif (RRM) is the most common RNA-binding domain in eukaryotes. Differences in RRM sequences dictate, in part, both RNA and protein-binding specificities and affinities. We used a deep mutational scanning approach to study the sequence-function relationship of the RRM2 domain of the Saccharomyces cerevisiae poly(A)-binding protein (Pab1). By scoring the activity of more than 100,000 unique Pab1 variants, including 1246 with single amino acid substitutions, we delineated the mutational constraints on each residue. Clustering of residues with similar mutational patterns reveals three major classes, composed principally of RNA-binding residues, of hydrophobic core residues, and of the remaining residues. The first class also includes a highly conserved residue not involved in RNA binding, G150, which can be mutated to destabilize Pab1. A comparison of the mutational sensitivity of yeast Pab1 residues to their evolutionary conservation reveals that most residues tolerate more substitutions than are present in the natural sequences, although other residues that tolerate fewer substitutions may point to specialized functions in yeast. An analysis of ∼40,000 double mutants indicates a preference for a short distance between two mutations that display an epistatic interaction. As examples of interactions, the mutations N139T, N139S, and I157L suppress other mutations that interfere with RNA binding and protein stability. Overall, this study demonstrates that living cells can be subjected to a single assay to analyze hundreds of thousands of protein variants in parallel.

Keywords: Pab1; RNA recognition motif; RNA-binding protein; epistasis; structure-function analysis.

FIGURE 1.

Experimental design of the deep mutational scan for the Pab1 RRM2 domain. (A) Protocol to assess the effects of RRM2 mutations on the in vivo function of Pab1. pab1Δ cells carry two plasmids, one expressing the full-length Pab1 protein under a tetracycline-off promoter (pTet-PAB1 WT) and the other expressing one of many variants of Pab1 from a constitutively active promoter (pGPD-PAB1 variant). The cells are grown to logarithmic phase in liquid culture, and a tetracycline analog (doxycycline) is added to the media. Cells expressing variants of Pab1 that cannot fully complement the loss of the wild-type protein grow slower than cells expressing neutral variants of Pab1. Sequencing the mutated fragments of the variant population before (input) and after selection (output) can be used to quantify these effects as the ratio of frequencies in each pool for each variant. (B) Selection of a Pab1 fragment that displays growth-rate sensitivity to a single point mutation. pab1Δ cells carrying the two plasmids specified in A with the variant plasmid expressing truncated forms of Pab1 with or without F170V substitution were tested for growth in the presence of doxycycline (20 μg/mL) on plates (top) and in liquid culture (bottom). Generation time was calculated starting from 8 h after doxycycline addition to eliminate cell divisions due to residual Pab1 activity. Note that cells carrying an empty vector grow at a low rate, probably due to leaky expression of the wild-type protein. (C) RRM2 mutagenesis. Shown are the Pab1(1–343BX) construct and the RRM2 sequence that was mutagneized. Each colored sequence and the corresponding elements on the structure of the human RRM2 domain (PDB_ID 1CVJ) represents a 25-amino acid-long RRM2 sequence that was doped with an average of three DNA base substitutions per variant. The graph at the bottom right depicts averages values of the listed properties of the three input libraries with respect to variants carrying a specified number of amino acid substitutions.

FIGURE 2.

Effect of single amino acid substitutions on the in vivo function of the Pab1 RRM2 domain. (A) A heat map displaying the enrichment scores (log₂ transformed) of single amino acid substitutions in the RRM2 domain. Each column represents an RRM2 sequence position and each row a substitution to a specific amino acid. An asterisk designates the row of nonsense mutations. Color ranges from blue for the most deleterious mutations to red for the most beneficial ones. Substitutions that were not sequenced in the input or selected pools or that were eliminated by subsequent quality filtration steps are shown in gray; wild-type residues are marked with white dots. The secondary structure of the RRM2 domain aligned to the sequence is shown below the heat map as well as the average enrichment score for each position. (B) The average enrichment scores are projected on the crystal structure of the human RRM2 (PDB_ID 1CVJ). RRM1 and the connecting linker are shown in black and the poly(A) RNA in orange.

FIGURE 3.

Clustering the effects of single amino acid substitutions groups structurally related residues. (A) Pab1 RRM2 positions and substituting amino acids were clustered based on enrichment score values and color coded as shown in Figure 2. The dotted line creates three clusters of RRM2 residues. Positions corresponding to RNA-binding residues in the human RRM2 are colored for their clustering (magenta) or lack of clustering (cyan) to group I. Positions corresponding to aliphatic residues are colored for their clustering (green) or lack of clustering (yellow) to group II. (B) Clustered residues displayed on the structure of the human RRM2 domain (PDB_ID 1CVJ) and color coded as in A. (Left) RNA-binding surface of RRM1–RRM2 with poly(A) shown in orange; (middle) RNA-binding residues with the poly(A) and the RNA-binding residues removed to observe the RRM2 core residues; (right) image as at left rotated 90° at the horizontal axis.

FIGURE 4.

Mutational sensitivity of residues in the helices α1 and α2 interface suggests a role for these residues in Pab1 stability. (A) Three of the four residues highly sensitive to mutation but not RNA binding (F149, G150, and F173) are found in close proximity in the RRM2 opening between the two helices. (B) Cold-suppressible phenotype of mutants carrying G150T, F149T, and F173A. pab1Δ cells carrying the two plasmids shown in Figure 1A with the variant plasmid expressing the specified mutations from Pab1(1–343BX) were grown in the absence or in the presence of 5-fluoro-orotic acid (5FOA) to follow the survival of the mutants upon wild-type plasmid loss. (C) Protease sensitivity of a G150T mutant. Western blot showing GST fusions of Pab1(1–343BX) constructs following treatment with increasing concentrations of proteinase K. (D) Conservation scores for multiple sequence alignment positions created from 119 RRM sequences in the protein data bank. The RNP elements and the four mutation sensitive residues are shown.

FIGURE 5.

Discrepancy between mutation sensitivity data and evolutionary conservation provides functional insights. (A, top) Logo plots generated for all amino acid substitutions that resulted in no more than a 5% reduction in performance compared with the wild type. Presented are only the four β-strands and flanking residues. (Bottom) Logo plots generated for the same RRM2 elements from a multiple sequence alignment of 306 Pab1 homologous sequences. The yeast RRM2 sequence corresponding to these logo plots is shown below. Residues shown to bind RNA in the human RRM2 domain (Deo et al. 1999) are underlined. (B) Comparison of the property entropy of each Pab1 RRM2 position in the multiple sequence alignment created from homologous sequences to the property entropies that were derived from all amino acid substitutions that showed no more than a 5% reduction in performance. The trendline is shown in a solid line, together with the Pearson's R² value. Dotted line represents perfect correlation. (C) The ratio between the functional conservation score to the evolutionary conservation score is color coded on the structure of the human RRM2 (PDB ID 1CVJ). From top, left in clockwise direction: lateral (facing RRM3), lateral (facing RRM1), dorsal, and ventral views of RRM2. RRM1 is shown in transparent green and the poly(A) in gray.

FIGURE 6.

Epistatic interactions in double mutants. (A) Contribution of spacing between residues in the primary sequence corresponding to the two mutations in each double mutant to epistatic interactions. A sequence distance of 0 corresponds to adjacent residues. (B) Contribution of physical distance between the two mutations forming each double mutant to epistatic interactions. Shown is the distribution of distances between the center of the masses of the two mutated residues based on the human RRM2 structure (PDB ID 1CVJ). Only variants with five or more residues separating the two mutated residues in the RRM2 sequence were included. For A and B, the P-values of wilcoxon-rank tests for the differences between the positive and negative epistasis groups to the no-epistasis group are specified. (C) Arc diagrams displaying the interactions between mutation pairs in variants with positive (red) or negative (blue) epistatic interactions. The sequence and the secondary structure of each segment are shown. The color of each node represents substitution to a specific amino acid and the size represents the fraction of variants with that particular mutation. A color map describing the identity of substituting amino acids is shown. (D) Presumed effects of N139 and I157 substitutions on strand β2 and helix α1 association may account for suppression of deleterious mutations. (Top, left) Structure of the human RRM2 in green with N112 and V130 residues shown in color-coded sticks (carbon, light blue; nitrogen, blue; oxygen, red). The hydrogen bond between the carbonyl group of the asparagine side chain and the amino group of the valine backbone is shown by a black dotted line, along with the distance. Hydrophobic residues found in close proximity to Val130 are also shown. (Top, right) Replacing the specified residues with the yeast residues supports a similar association between I157 and N139 and between I157 to the conserved aliphatic residues. (Bottom, left) Substitution N139T (as well as N139S) may weaken the hydrogen bonding with I157 by increasing the distance between the hydrogen donor and acceptor groups. (Bottom, right) Substitution I157L may cause a similar effect by destabilizing Loop L1 and L3 association.