Modulation of gurken translation by insulin and TOR signaling in Drosophila

Abstract

Localized Gurken (Grk) translation specifies the anterior-posterior and dorsal-ventral axes of the developing Drosophila oocyte; spindle-class females lay ventralized eggs resulting from inefficient grk translation. This phenotype is thought to result from inhibition of the Vasa RNA helicase. In a screen for modifiers of the eggshell phenotype in spn-B flies, we identified a mutation in the lnk gene. We show that lnk mutations restore Grk expression but do not suppress the persistence of double-strand breaks nor other spn-B phenotypes. This suppression does not affect Egfr directly, but rather overcomes the translational block of grk messages seen in spindle mutants. Lnk was recently identified as a component of the insulin/insulin-like growth factor signaling (IIS) and TOR pathway. Interestingly, direct inhibition of TOR with rapamycin in spn-B or vas mutant mothers can also suppress the ventralized eggshell phenotype. When dietary protein is inadequate, reduced IIS-TOR activity inhibits cap-dependent translation by promoting the activity of the translation inhibitor eIF4E-binding protein (4EBP). We hypothesize that reduced TOR activity promotes grk translation independent of the canonical Vasa- and cap-dependent mechanism. This model might explain how flies can maintain the translation of developmentally important transcripts during periods of nutrient limitation when bulk cap-dependent translation is repressed.

Fig. 1.

lnk mutations suppress dorsal–ventral patterning defects in eggs laid by spn-B flies. Mutations in spindle-class genes result in female sterility and variable ventralization of the eggs laid by homozygous females. (A) Scanning electron microscopy images of eggs laid by spn-B^BU flies demonstrating a range of ventralization phenotypes. Dorsal–ventral polarity was classified as wild type if two distinct dorsal appendages were evident. Fusion at the base of the appendages was classified as mild class 2 ventralization (V2), whereas fusion past the anterior aspect of the egg when viewed dorsally was classified as moderate class 3 ventralization (V3). Severely ventralized eggs lacking all appendage material were designated as V4. Scale bar: 100 μm. (B) Genetic suppression of the spn-B^BU ventralized eggshell phenotype by mutations in mei-P22^CA1215 and several allelic combinations with lnk^CR642. Mutations in lnk are also able to suppress dorsal–ventral patterning defects in eggs laid by spn-A^093/003 mutant females. (C) A domain map of the Lnk protein illustrating the location of the CR642 mutation and conserved tyrosine phosphorylation site. (D–F) The eggshell suppression reflects a rescue of Grk protein expression in the ovary. Grk is stained green; F-actin is shown in red. Scale bars: 25 μm.

Fig. 2.

Lnk functions in both the germline and follicle cells. (A) Immunostaining of Lnk protein (red in A and white in A′) in follicle cell mosaics generated in e22c-FLP/+; FRT82, lnk^CR642/FRT82, ubi-GFP flies. Clones are marked by the absence of EGFP and DNA was stained blue with Hoechst 33342. Images are a maximum intensity projection of multiple confocal sections. (B–D) Localization of an N-terminal EGFP–Lnk fusion construct expressed from the genomic promoter. B is a single confocal section through the germarium and early stage cysts, C is a section through a stage 10A cyst, and D shows a maximum intensity projection of multiple sections through the nurse cell cluster of a stage 10B egg chamber. (E) Expression of a genomic lnk rescue construct can restore a ventralized eggshell phenotype to spn-B^BU, lnk^CR642 flies. (F) nos-Gal4 or CY2-Gal4 were used to drive expression of a UAS-lnk construct in the germline or follicle cells respectively of spn-B^BU, lnk^CR642 flies. Act 5C-Gal4 was used to drive expression in both tissues. Germline expression of lnk resulted in a significant increase in the number of ventralized eggs whereas follicle cell expression did not. Eggshell classes are as described in Fig. 1. Scale bars: 50 μm (A,C,D), 25 μm (B).

Fig. 3.

lnk mutations do not affect Egfr activity in the follicle cells. (A,B) Immunostaining of follicle cell clones in stage 10 e22c-FLP/kek-LacZ; FRT82, lnk^CR642/FRT82, ubi-GFP ovaries. β-galactosidase (red) expression from the kek-LacZ enhancer trap reflects the gradient of Egfr activity in the follicle cells. Homozygous lnk^CR642 cells are marked by the absence of EGFP. A is a maximum intensity projection of multiple confocal sections encompassing the peak of Egfr activity at the future dorsal anterior that declines towards the ventral and posterior axes; B is a single confocal section with the dorsal anterior on the top left. kek-LacZ expression is unaffected by the lnk genotype. Scale bars: 50 μm. (C) Cbl binds to the conserved tyrosine residue at 720 in SH2-B family members. Mutation of Lnk tyrosine 720 to a non-phosphorylatable phenylalanine residue in a genomic rescue construct does not affect the ability of the transgene to restore a ventralized eggshell phenotype in spn-B^BU/spn-B^Δ37C, lnk^CR642/lnk^f05062 flies. Eggshell classes are as described in Fig. 1.

Fig. 4.

lnk mutations affect IIS and slow the rate of oogenesis. Whole ovaries were imaged from Oregon R (A,D), lnk^d07478 (B,E) and lnk^CR642 (C,F) flies. Ovaries in D–F were stained with Draq5 and imaged with UV epi-illumination. Following the onset of vitellogenesis, the oocyte autofluoresces blue. Note the strong arrest in early vitellogenesis in lnk^d07478 ovaries, whereas many late stage egg chambers are evident in lnk^CR642 ovaries. (G) Membrane localization of the tGPH reporter is sensitive to PtdIns(3,4,5)P₃ levels. A follicle cell clone from tGPH /+; FRT82, lnk^f02642/FRT82 ovaries stained with anti-Lnk antibodies (red, and G′) illustrating the reduced membrane localization of tGPH (green, G″) in cells lacking Lnk activity. The clone border is illustrated by the dotted line with heterozygous tissue on the left and lnk^f02642 homozygous cells on the right. (H) Linage tracing in wild-type and lnk^CR642 ovaries using the hsFLP X15 system. Clones were induced in the germarium by transient heat shock and allowed to develop on plates with live yeast paste for 4 days. The latest stage egg chamber with multiply marked nurse cell nuclei in each ovariole was recorded for each genotype. lnk^CR642 egg chambers developed more slowly than wild type. Scale bars: 500 μm (A–F), 10 μm (G).

Fig. 5.

lnk suppression does not affect DNA repair or vas phenotypes. (A–D) Immunostaining of γ-H2Av foci (green) in germaria. Orb (red) stains individual cysts in the germarium and accumulates in the oocyte in region 2B. DNA is shown in blue. γ-H2Av foci appear in region 2A of wild-type germaria and are resolved in oocytes in region 2B and 3 (arrowheads) whereas DSBs are evident in oocytes between region 2B and stage 4 in spn-B^BU ovarioles. Although spn-B^BU, mei-P22^CA1215 and spn-B^BU, lnk^CR642 flies both lay wild-type eggs, only mei-P22^CA1215 does so by blocking the formation of DSBs. spn-B^BU, lnk^CR642 germaria have persistent DSBs in region 3 to a similar extent as those in spn-B^BU. (E–H) A single confocal section of the karyosome from mid-stage egg chambers stained with Hoechst 33342 reveals that the karyosome fails to properly condense in spn-B^BU, lnk^CR642 oocytes and resembles the phenotype seen in spn-B^BU. (I) Western blot of Vasa protein from ovarian extracts reveals that the electrophoretic mobility of Vasa from spn-B^BU, lnk^CR642 retains the modification seen in spn-B^BU. Scale bars: 5 μm (A–D), 1 μm (E–H).

Fig. 6.

Suppression of spn-B and vas eggshell phenotypes by rapamycin. (A) spn-B^BU and (B) vas^PH165/RG53 flies were fed yeast paste on grape juice agar containing the indicated concentration of rapamycin. Eggs from days 4–7 were scored for eggshell phenotypes. In both cases, TOR inhibition results in strong suppression of eggshell ventralization. At 10 μM rapamycin completely inhibited egg deposition in spn-B^BU flies. Eggshell classes are as described in Fig. 1.

Fig. 7.

A model of Vasa-independent grk translation. (A) When nutrients are abundant, cap-dependent translation predominates. Vasa helicase activity facilitates 43S PIC scanning of the 5′-UTR, allowing it to navigate secondary structures. Once an AUG codon is identified, Vasa–eIF5B interactions promote joining of the 60S subunit to form the 80S ribosome. (B) When nutrients are low, IIS is compromised or flies are fed the TOR-C1 inhibitor rapamycin, TOR activity falls and 4EBP is free to inhibit the cap-binding protein eIF4E. The resulting increase in free ribosomes favors IRES translation initiation of the localized grk transcript allowing oocytes to be patterned correctly, thereby preserving their viability for when nutrients are available again. (C) spindle-class mutants such as spn-B^BU inhibit grk translation by phosphorylating Vasa in a checkpoint-dependent manner. This disrupts scanning of the 43S PIC and/or interactions with eIF5B (5B) and therefore subunit joining at the initiation AUG codon. In these conditions, grk cannot be translated and ventralized eggs result. (D) Reduced TOR activity in a spn-B mutant background permits grk translation independent of Vasa activity. Despite the inhibition of cap-dependent translation initiation resulting from Vasa phosphorylation and increased 4EBP activity, grk translation persists. We hypothesize that this is due to alternative translation initiation at an IRES in the grk 5′-UTR.