Fertility and polarized cell growth depends on eIF5A for translation of polyproline-rich formins in Saccharomyces cerevisiae

Abstract

eIF5A is an essential and evolutionary conserved translation elongation factor, which has recently been proposed to be required for the translation of proteins with consecutive prolines. The binding of eIF5A to ribosomes occurs upon its activation by hypusination, a modification that requires spermidine, an essential factor for mammalian fertility that also promotes yeast mating. We show that in response to pheromone, hypusinated eIF5A is required for shmoo formation, localization of polarisome components, induction of cell fusion proteins, and actin assembly in yeast. We also show that eIF5A is required for the translation of Bni1, a proline-rich formin involved in polarized growth during shmoo formation. Our data indicate that translation of the polyproline motifs in Bni1 is eIF5A dependent and this translation dependency is lost upon deletion of the polyprolines. Moreover, an exogenous increase in Bni1 protein levels partially restores the defect in shmoo formation seen in eIF5A mutants. Overall, our results identify eIF5A as a novel and essential regulator of yeast mating through formin translation. Since eIF5A and polyproline formins are conserved across species, our results also suggest that eIF5A-dependent translation of formins could regulate polarized growth in such processes as fertility and cancer in higher eukaryotes.

Keywords: eIF5A; formins; mating; polarized growth; translation.

Figure 1

Number and functional categories of S. cerevisiae proteins containing three or more consecutive prolines. (A) Number of yeast proteins containing polyPro stretches with the indicated consecutive prolines. Formins Bni1 and Bnr1 are the only yeast proteins with polyproline stretches with 12 or more prolines. (B) GO functional categories significantly overrepresented in the 549 yeast genes containing a polyPro stretch with at least three consecutive prolines. Bars represent the number of genes found in each category and the P-value is indicated for each category.

Figure 2

Yeast shmoo formation requires hypusinated eIF5A. (A and B) Representative DIC images of wild type and mutants with or without treatment of 10 μg/ml α-factor for 2 hr. Null mutants were maintained in 30° (A), and temperature-sensitive mutants were maintained at 25° or transferred to 37° for 4 hr (B). Quantification of percentage of cells containing shmoos in samples treated with α-factor are indicated. Approximately 300 cells were manually counted for each sample from at least two independent experiments.

Figure 3

Localization of the polarisome protein Spa2 at the shmoo tip in response to pheromone requires eIF5A. Representative GFP and their corresponding DIC images of wild type, bni1Δ, tif51A-1, and tif51A-3 mutants overexpressing Spa2-GFP. Cells were maintained at 25° or transferred to 37° for 4 hr. Cells were examined after treatment with 10 μg/ml α-factor for 2 hr.

Figure 4

Expression of FUS1-lacZ in response to pheromone requires eIF5A. (A and B) Expression of FUS1-LacZ (A) or STL1-LacZ fusion (B) in wild type and mutants. Cells were grown at 25° or 4 hr at 37° and then treated with 10 μg/ml α-factor for either 1 or 2 hr (A) or stressed with 0.6 M KCl for 45 min (B). Error bars represent the standard error of the mean of three or more biological replicates.

Figure 5

Polymerization and localization of actin cables require eIF5A during shmoo formation. Representative fluorescence and their corresponding DIC images of wild type, tif51A-1, tif51A-3, and bni1Δ showing F-actin staining by phalloidin at 25° or after incubation for 4 hr at 37°, in the presence or absence of treatment with 10 μg/ml α-factor for 2 hr.

Figure 6

Translation of the proline stretches of the formin BNI1 requires eIF5A. (A) Schematic diagrams showing C-terminal HA genomic tagging of full-length (BNI1-HA), C-terminal truncated (Δ1240–1954, BNI1ΔCt-HA), and proline-deleted (Δ1239–1307, BNI1ΔPro-HA) Bni1. Formin homology (FH) and protein interaction domains are indicated, as well as the polyPro stretches with the number of consecutive prolines represented in red. The numbers below the boxes indicate the first amino acid of each domain or stretch. (B) Immunoblots with antibodies against HA and hexokinase 2 protein (Hxk2) to show expression of full-length Bni1, C-terminal truncated Bni1, proline-deleted Bni1 and Hxk2, in wild type and tif51A-1 at 25° or 37° at the indicated times. (C) Translation efficiency of BNI1-HA, BNI1ΔCt-HA, and BNI1ΔPro-HA relative to translation efficiency of HXK2 in wild type and tif51A-1. Protein/mRNA ratios for Bni1, Bni1ΔCt-HA, Bni1ΔPro-HA, and Hxk2 were calculated by Western (B) and quantitative PCR from the same samples. Translation efficiencies of BNI1-HA, BNI1ΔCt-HA, and BNI1ΔPro-HA are calculated relative to translation efficiency of HXK2 and represented as a fraction against 25° for each strain.

Figure 7

Overexpression of the formin BNI1 partially restores the shmoo formation defect of eIF5A mutants. (A) Overexpression of BNI1 does not restore the growth defect of eIF5A thermosensitive mutants at restrictive temperature. Growth of the indicated yeast strains containing a plasmid expressing BNI1 under the GAL promoter (pGAL-BNI1) was tested in Sgal −ura plates incubated at the indicated temperatures. (B) Immunoblots with antibodies against HA and hexokinase 2 protein (Hxk2) to show the expression of genomic Bni1 (BNI1) or the overexpression of Bni1 (pGAL-BNI1) at 25° or 37° for 4 hr. (C) Shmoo formation is partially restored in eIF5A mutants. The indicated strains containing a pGAL-BNI1 plasmid were grown at 25° in Sgal−ura media and then were maintained at 25° or transferred to 37° for 4 hr. Quantification of percentage of cells forming shmoos in samples treated with 10 μg/ml α-factor are indicated. Approximately 300 cells were manually counted for each sample from at least two independent experiments.