Evolutionary conserved role of eukaryotic translation factor eIF5A in the regulation of actin-nucleating formins

Abstract

Elongation factor eIF5A is required for the translation of consecutive prolines, and was shown in yeast to translate polyproline-containing Bni1, an actin-nucleating formin required for polarized growth during mating. Here we show that Drosophila eIF5A can functionally replace yeast eIF5A and is required for actin-rich cable assembly during embryonic dorsal closure (DC). Furthermore, Diaphanous, the formin involved in actin dynamics during DC, is regulated by and mediates eIF5A effects. Finally, eIF5A controls cell migration and regulates Diaphanous levels also in mammalian cells. Our results uncover an evolutionary conserved role of eIF5A regulating cytoskeleton-dependent processes through translation of formins in eukaryotes.

Figure 1

Knockdown of Dm eIF5A in the embryonic epidermis produces defects in DC and actin cable assembly. (A) Schematic representation of the anterior-posterior (AP) compartment boundary of a Drosophila wing imaginal disc (shaded in pink). (B) Anti-eIF5A antibody staining of a dpp-GAL4/UAS-iReIF5A wing imaginal disc, in which eIF5A knockdown is induced in the AP boundary. (C,D) Dark-field micrographs of representative (C) wild-type and (D) eIF5A mutant embryos (eIF5A ^KD). Anterior is to the left, dorsal is up. White arrowheads point to holes or not properly sealed areas in the anterior and dorsal parts of the embryo. (E,F) Confocal images of the lateral epidermis, DME cells and AS cells of representative stage 13–14 (E) control and (F) eIF5A ^KD embryos stained with phalloidin. Blue arrows point to the actomyosin cable in DME cells, which is irregular and discontinuous in eIF5A mutants. Detachment of DME cells from AS cells is frequently observed in mutant embryos (red arrowheads in F). Scale bars: 100 µm in C,D; 25 µm in E,F.

Figure 2

Heterologous expression of Dm eIF5A restores growth and shmoo formation of Sc eIF5A mutant cells. (A) Wild-type and tif51A-1 or tif51A-3 S. cerevisiae mutants containing empty plasmid or plasmids expressing Dm eIF5A or Sc eIF5A genes were grown at permissive temperature (25 °C) until exponential phase and then plated (10-fold serial dilutions) onto YPD medium and incubated at the indicated temperatures. (B) Same yeast strains and transformants described in (A) were grown until exponential phase and then treated or not with 10 μg/ml α-factor for 2 h. Yeast cells were then maintained at 25 °C or transferred to 37 °C for 4 h and DIC images were reported. Quantifications of percentage of cells containing shmoos in samples treated with α-factor are indicated. Approximately 200 cells were manually counted for each sample from at least two independent experiments.

Figure 3

Heterologous expression of Dm eIF5A allows translation of the yeast polyPro formin BNI1 in Sc eIF5A-depleted cells. (A) Schematic diagrams with domains of HA genomic-tagged Bni1. The FH domains, the polyPro stretches with the number of consecutive prolines (red) and the position of the first amino acid of each domain/stretch (below) are indicated. (B) Western blot for Bni1-HA (anti-HA) and Hxk2 expression in wild-type and tif51A-1 cells at 25 or 37 °C at the indicated times. (C) Translation efficiency of BNI1-HA relative to that of HXK2 in wild-type and tif51A-1 cells. Protein/mRNA ratios for Bni1 and Hxk2 were calculated by Western and qRT-PCR from same samples. Translation efficiency of BNI1-HA was calculated relative to translation efficiency of HXK2 and represented as a fraction against 25 °C for each strain and from two independent experiments. Data are represented as mean ± SD. Two-tailed student’s t-test analysis: *p < 0.05, **p < 0.01.

Figure 4

Reduction of Dm eIF5A function affects Dia formin levels. (A,B) Confocal images of a representative stage 13 en-GAL4>UAS-GFP; UAS-iReIF5A (en>GFP>iReIF5A) embryo stained with anti-Dia. A lateral view of the epidermis is shown. Anterior it to the left, dorsal is up. (A) Overlay of the en-GAL4 positive bands, identified by GFP expression, and Dia localization. (B) Image of the anti-Dia staining in which yellow lines represent the limits of the en-GAL4 positive bands. Decreased Dia levels are observed in cells with reduced eIF5A function (GFP marked cells) when compared to nearby non-GFP wild-type cells. Scale bar: 25 µm. (C) Quantification of the anti-Dia signal intensity (fluorescence intensity) in eIF5A mutant cells (GFP) with respect to control cells (Non-GFP). Data are represented as mean ± SD. Two-tailed student’s t-test analysis: *p < 0.05.

Figure 5

Cell migration and Dia levels are regulated by eIF5A in mouse NSCs. (A) Immunocytochemistry for hypusine and DAPI staining in migrating neurospheres with or without 10 μM GC7. (B) Western blot for hypusinated eIF5A (FabHpu24) and control GAPDH in NSCs treated during 3 days with GC7 (5 and 10 μM) and then washed and cultured for 3 more days (5w and 10w) (n = 3). (C) Quantification of the migration area of individual neurospheres 7 h after treatment (32 to 44 neurospheres analyzed). (D) Representative images of the neurosphere migration assay. Migration front and original neurosphere are marked with dashed lines. (E) Immunocytochemistry for mDIA1 and DAPI staining in untreated and GC7 treated neurospheres in migration conditions. (F) Western blot for mDIA1 and control α-TUBULIN in NSCs treated for 3 days and then washed (n = 3). Two-tailed Student’s t-test analysis: *p < 0.05, **p < 0.01, ***p < 0.001. Scale bars: A and E, 10 μM; D, 100 μm.