Functional specificity among ribosomal proteins regulates gene expression

Abstract

Duplicated genes escape gene loss by conferring a dosage benefit or evolving diverged functions. The yeast Saccharomyces cerevisiae contains many duplicated genes encoding ribosomal proteins. Prior studies have suggested that these duplicated proteins are functionally redundant and affect cellular processes in proportion to their expression. In contrast, through studies of ASH1 mRNA in yeast, we demonstrate paralog-specific requirements for the translation of localized mRNAs. Intriguingly, these paralog-specific effects are limited to a distinct subset of duplicated ribosomal proteins. Moreover, transcriptional and phenotypic profiling of cells lacking specific ribosomal proteins reveals differences between the functional roles of ribosomal protein paralogs that extend beyond effects on mRNA localization. Finally, we show that ribosomal protein paralogs exhibit differential requirements for assembly and localization. Together, our data indicate complex specialization of ribosomal proteins for specific cellular processes and support the existence of a ribosomal code.

Figure 1

Loc1 is required for the translational regulation of ASH1 mRNA. (A) Reporter constructs used to assay ASH1 regulation. The reporters contain the promoter and ORF of yeast PGK1, an array of U1A hairpins, and either PGK1’s own 3’ UTR or one of ASH1’s four localization elements. Each reporter was co-expressed along with U1A-GFP, which specifically binds the U1A hairpins and allows visualization of reporter mRNA location in live cells. The fraction of cells with either bud-tip, bud-cytoplasm, or ubiquitous localization was determined (see Experimental Procedures). (B) Defective anchoring of the E3 construct in loc1 Δ cells is due to aberrant translation. Histograms of the E3 reporter construct localizations in wild-type, loc1 Δ, and wild-type cells following brief treatment with cycloheximide. Error bars represent total variation between replicate experiments. (C) Protein level of myc-Ash1 increases in loc1 Δ cells relative to actin (negative control). Western blots of myc-Ash1 and actin were performed from extracts of wild-type and loc1 Δ cells. Equal amounts of protein were loaded in each lane. (D) mRNA level of ASH1 decreases in loc1 Δ cells.

Figure 2

Genes required for bud-site selection in yeast are also required for ASH1 localization. (A) Representative images of cells expressing ASH1 reporter. (B) Strains that have defective bud-site selection also have defects in localization of the E3 reporter construct. Fraction of cells exhibiting bud-tip, bud-cytoplasm, bud-neck, and “other” (not bud-tip, bud-neck, or bud-cytoplasm) localizations of the E3 reporter construct in cultures lacking the genes indicated is shown. Error bars represent standard deviations of replicate experiments.

Figure 3

Regulated translation of the E3 reporter construct requires a specific subset of duplicated ribosomal protein genes. (A) – (B), Error bars represent standard deviations of replicate experiments. (A) Ribosomal proteins that are required for bud-site selection have a larger defect in anchoring of the E3 reporter construct than their nearly-identical paralogs. Fraction of cells exhibiting either bud-tip or bud-cytoplasmic localization of the E3 reporter construct in cells lacking the gene is indicated. Genes that are required for bud-site selection in diploids are indicated by “*”. (B) There is a significantly greater difference in the effect on anchoring of the E3 reporter construct between pairs of duplicated ribosomal protein genes in which one copy is required for bud-site selection than for pairs in which neither copy is required for bud-site selection. The fraction of cells exhibiting bud-tip and bud-cytoplasmic localization of the E3 reporter construct was assayed in strains lacking a variety of duplicated ribosomal protein genes. The difference between the fraction of cells exhibiting bud-tip localization is plotted against the difference in the fraction exhibiting bud-cytoplasmic localization for both members of each pair.

Figure 4

Phenotypic data reveals complex functional relationships between duplicated ribosomal protein genes. (A) Hierarchical clustering analysis of phenotypic data by ribosomal protein (vertical axis) and phenotype (horizontal axis). Although many ribosomal proteins shared some phenotypes, no two proteins are required for the same set of processes, and different groups are required for each process. (B) Paralogous ribosomal proteins are not phenotypically similar. Rpl2a and Rpl2b cluster with completely different groups of genes, as indicated by the shaded boxes that correspond to (A). (C) Paralogous ribosomal proteins share no more phenotypes than non-paralogous genes. The number of shared phenotypes between all combinations of duplicated ribosomal protein genes was calculated and sorted into paralogous or non-paralogous relationships. Normalized values are displayed. (D) Phenotypic effects are not determined by expression level. mRNA expression levels of all duplicated ribosomal protein genes from transcriptional profiling data was used to determine the relative contribution of each paralog. Genes were sorted into “higher” or “lower” based on whether they contributed more or less than half of the mRNA, respectively. Error bars represent standard deviations.

Figure 5

Paralogous ribosomal proteins exhibit different localizations and assembly requirements in specific genetic backgrounds. GFP-tagged Rpl7a, Rpl7b, Rps18a, and Rps18b were expressed from the genome under their own promoters in wild-type, loc1Δ , puf6Δ, and loc1Δpuf6Δ cells. Representative fluorescent (top) and nomarski (bottom) images are shown.