Mextli is a novel eukaryotic translation initiation factor 4E-binding protein that promotes translation in Drosophila melanogaster

Abstract

Translation is a fundamental step in gene expression, and translational control is exerted in many developmental processes. Most eukaryotic mRNAs are translated by a cap-dependent mechanism, which requires recognition of the 5'-cap structure of the mRNA by eukaryotic translation initiation factor 4E (eIF4E). eIF4E activity is controlled by eIF4E-binding proteins (4E-BPs), which by competing with eIF4G for eIF4E binding act as translational repressors. Here, we report the discovery of Mextli (Mxt), a novel Drosophila melanogaster 4E-BP that in sharp contrast to other 4E-BPs, has a modular structure, binds RNA, eIF3, and several eIF4Es, and promotes translation. Mxt is expressed at high levels in ovarian germ line stem cells (GSCs) and early-stage cystocytes, as is eIF4E-1, and we demonstrate the two proteins interact in these cells. Phenotypic analysis of mxt mutants indicates a role for Mxt in germ line stem cell (GSC) maintenance and in early embryogenesis. Our results support the idea that Mxt, like eIF4G, coordinates the assembly of translation initiation complexes, rendering Mxt the first example of evolutionary convergence of eIF4G function.

Fig 1

Mxt is a novel type of eIF4E-interacting protein that binds RNA, eIF4Es, eIF3, and CG3225. (A) Schematic representation of the domain structure and interactions of Mextli, Saccharomyces cerevisiae (S.) eIF4G1, and human (H.) eIF4G. The MIF4G (amino acids 1 to 130), KH RNA binding domains (amino acids 240 to 279), and the eIF4E-binding motif (amino acids 581 to 587) of Mxt are indicated. (B) Mxt interacts with eIF4E-1, -2, -3, -4, and -7 in the yeast two-hybrid system via the eIF4E-binding site. -L, −Leu; -W, −Trp; -A, −Ade; -H, −His. (C to G) Coimmunoprecipitation experiments showing that Mxt physically interacts with eIF4E-1, eIF3, and the helicase encoded by the annotated gene CG3225. eIF3e, eIF3h, and CG3225 are in fusion with the V5 epitope. Plasmids expressing 3×HA-tagged versions of Mxt (C and D), Mxt^AAA (E), GFP (negative control) (F), or eIF4E-1 (G) were transfected into S2 cells. For a schematic representation of the constructs used (3×HA-Mxt and 3×HA-Mxt^AAA), see Fig. 2A. (G) HA-eIF4E-1 physically interacts with endogenous Mxt, eIF4G, and 4E-BP (Thor). Immunoprecipitations (IP) were conducted using either beads alone or beads plus anti-HA antibodies in the absence (C) or presence (D, E, F, and G) of RNase A, and interactions were detected by immunoblotting.

Fig 2

Mapping of Mxt regions for interaction with eIF4E-1 and eIF3. (A) Plasmids that express different 3×HA-tagged regions of Mxt in S2 cells. The 3×HA-Mxt and 3×HA-Mxt^AAA constructs were used in Fig. 1C to E. (B) The plasmids depicted in panel A were either transfected into S2 cells alone or cotransfected with a plasmid expressing a V5-tagged version of eIF3h. Total extracts of transfected cells were used to perform immunoprecipitation experiments using either beads alone or beads plus anti-HA antibodies in the presence of RNase A. Samples were then subjected to Western blotting (WB) using either anti-eIF4E-1 or anti-V5 antibodies. i, input.

Fig 3

Mxt promotes translation. (A) m⁷GTP-Sepharose pulldown. Plasmids expressing 3×HA-tagged versions of Mxt or Mxt^AAA were transfected into S2 cells, and 50 μg of total extract of transfected cells was subjected to m⁷GTP-Sepharose chromatography. Bound and unbound fractions were then subjected to SDS-PAGE and Western blotting using anti-HA or anti-eIF4E-1 antibodies. i, input. (B) Luciferase assays measuring translation of an m⁷G-capped and polyadenylated firefly luciferase reporter (FLuc) mRNA in cell-free ovary lysates from yw (wild-type) or homozygous mxt^e00436 (mxt^−/−) females. The bars represent the results of five independent experiments (P < 0.001). (C) Coomassie blue staining of recombinant 6×His-Mxt protein. Lane M, molecular mass markers (kDa). (D) Luciferase assays measuring translation in homozygous mxt^e00436 (mxt^−/−) ovary lysates supplemented with either buffer or 0.2 μg of recombinant 6×His-Mxt. *, P < 0.001.

Fig 4

Structure and expression of the mxt gene. (A) Schematic diagram of the mxt gene (CG2950) showing predicted transcripts and sites of the PiggyBac insertions f01505 and e00436. (B) Northern analysis shows that mxt mRNA is ∼3.7 kb in length. (C) Western blot analysis shows that affinity-purified anti-Mxt antibodies (2101) recognize a single polypeptide band of the predicted molecular mass of Mxt (70 kDa) in total extracts (15 μg per lane) that is greatly reduced in mxt mutants. Similar results were obtained with affinity-purified anti-Mxt antibodies 2103 and 2104 (data not shown). α-Tub, α-tubulin.

Fig 5

Maternal loss of mxt function affects embryogenesis. (A) Hatching test. Females or males of the indicated phenotype were mated with wild-type (WT) (Oregon-R) animals of the opposite sex, and eggs that failed to hatch were counted. Hatching rates were significantly lower among embryos produced by mxt females than among those from wild-type females. (B) Cuticle preparations from wild-type embryos and embryos that were produced by mxt^e00436/Df(2L)Exel6010 females and failed to hatch. Thoracic (T1 to T3) and abdominal (A1 to A8) segments, mouth apparatus (MA), and the denticle belts (*) are labeled in the wild-type embryo. Embryos from mxt females exhibit various segmentation defects, including deletion of one or more thoracic or abdominal segments (left panels), severe global defects (middle panels), or malformation or incomplete development of abdominal segments (right panels). (C) In situ hybridization experiments for ftz mRNA in embryos produced by females of the indicated genotypes. The seven ftz stripes in the wild-type embryo are labeled and numbered. mxt mutants showed partial or total deletion of the fourth stripe and/or fusion of sixth and seventh stripes or widening of the seventh stripe. These defects were found in 25 to 32% of embryos produced by mutant females, as indicated.

Fig 6

mxt mutations affect germarium morphology and female fertility. (A) Schematic diagram of the Drosophila germarium. Germ line stem cells (GSCs) divide asymmetrically to form another GSC and a cystoblast, which will subsequently divide to give rise to 16 germ line cells, one of which will differentiate into the oocyte. The remainder will become nurse cells. Each GSC and cystoblast contains a spectrosome. During transit-amplifying divisions, the spectrosome develops into a branched structure called the fusome, which extends through cytoplasmic bridges through each germ line cell and orients the axes of the cell divisions (62). (B) Immunostaining of ovaries from 20-day-old females of the indicated genotypes with Mei-P26 (red), Orb (green), and the nuclear stain DAPI (4′,6-diamidino-2-phenylindole) (blue). (C) Egg-laying tests of mxt^e00436/Df(2L)Exel6010 (green bars) and +/Df(2L)Exel6010 (red bars; control) females. mxt^e00436/Df(2L)Exel6010 females produce significantly fewer eggs (P < 0.005 after 15 days of aging) over time than control females. (D) Immunostaining of germaria from 21-day-old mxt^e00436/CyO (436/CyO) and mxt^e00436/Df(2L)Exel6010 (436/Df) ovaries with anti-Add87/Hts (1B1) antibodies. Spectrosomes (ss) and branched fusomes (f) are labeled. mxt^e00436/Df(2L)Exel6010 germaria contained fewer spectrosomes, which are present in germ line stem cells and cystoblasts, than the heterozygous control.

Fig 7

Mextli and eIF4Es expression during oogenesis. (A) Composites of multiple images depicting immunostaining of wild-type (Oregon-R) or mxt^e00436/Df(2L)Exel6010 ovarioles (negative control) with anti-Mxt antibodies. Mxt is cytoplasmic and enriched in the GSCs and cystoblasts and in the perinuclear nuage. Somatic expression in stretched cells and columnar follicular cells is also apparent. Arrows indicate sites of high Mxt concentration. The right panels show magnified stage (St.) 10 (upper panel) and stage 4 (lower panel) egg chambers with enrichment of Mxt (arrows) in the cytoplasm of a stretched cell and in the perinuclear region of nurse cells, respectively. (B) Composites of multiple images depicting immunostaining of wild-type ovaries with anti-eIF4E-1 or anti-eIF3j antisera. Note the enrichment of eIF4E-1 in the GSCs and cystoblasts (white arrow). (C) Immunostaining of wild-type germaria with anti-Mxt and anti-Mei-P26 antisera. Mxt expression is high in GSCs and cystocytes but reduced in transit-amplifying cells, where Mei-P26 expression peaks. (D) Immunostaining of mei-P26^fs1/mei-P26^mfs1 ovaries with anti-Mxt and anti-Mei-P26 antisera. Mxt expression levels are heterogeneous among undifferentiated germ cells within individual tumorous egg chambers.

Fig 8

Mxt and eIF4E-1 colocalize and physically interact in germ line stem cells and during oogenesis. (A to C) Mxt and eIF4E-1 are enriched and colocalize in GSCs and cystoblasts. (A) Coimmunostaining of wild-type (Oregon-R) germarium with either anti-eIF4E-1 (green) and anti-ADD87-Hts (1B1; red) or anti-Mxt (red) and anti-ADD87-Hts (1B1; green). eIF4E-1 and Mxt are enriched in GSCs and cystoblasts, as marked by punctate staining with 1B1 antibody. (B) Immunostaining of germaria from GFP-eIF4E-1/GFP-eIF4E-1 flies with anti-Mxt and anti-GFP showing results similar to those in panel A. (C) Immunoprecipitation experiments using total extracts of bam^−/− germaria (made up mostly by GSCs) performed using either anti-eIF4E-1 antibodies or preimmune serum in the presence of RNase A showing that endogenous eIF4E-1 and Mxt can be copurified. (D) Immunoprecipitation experiments using total extracts of wild-type (Oregon-R) ovaries, again showing an association between endogenous Mxt and eIF4E-1 (left), as well as between Mxt and eIF3b (right). (E) Immunoprecipitation experiments using total extracts of ovaries either from wild-type (Oregon-R, control) or transgenic females expressing V5-tagged versions of Mxt or Mxt^AAA. Binding of Mxt to eIF4E-1 requires an intact consensus eIF4E-binding domain.